High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough,...

15
Ecology and Evolution. 2018;1–15. | 1 www.ecolevol.org Received: 28 June 2018 | Revised: 10 October 2018 | Accepted: 15 October 2018 DOI: 10.1002/ece3.4688 ORIGINAL RESEARCH High genomic diversity and candidate genes under selection associated with range expansion in eastern coyote ( Canis latrans ) populations Elizabeth Heppenheimer 1 | Kristin E. Brzeski 1,2 | Joseph W. Hinton 3 | Brent R. Patterson 4,5 | Linda Y. Rutledge 1,5 | Alexandra L. DeCandia 1 | Tyler Wheeldon 4,5 | Steven R. Fain 6 | Paul A. Hohenlohe 7 | Roland Kays 8,9 | Bradley N. White 5 | Michael J. Chamberlain 3 | Bridgett M. vonHoldt 1 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2018 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. 1 Department of Ecology & Evolutionary Biology, Princeton University, Princeton, New Jersey 2 School of Forest Resources and Environmental Science, Michigan Technological University, Houghton, Michigan 3 Warnell School of Forestry and Natural Resources, University of Georgia, Athens, Georgia 4 Ontario Ministry of Natural Resources and Forestry, Peterborough, Ontario, Canada 5 Trent University, Peterborough, Ontario 6 USFWS National Forensics Laboratory, Ashland, Oregon 7 Department of Biological Sciences, Institute for Bioinformatics and Evolutionary Studies, University of Idaho, Moscow, Idaho 8 Department of Forestry and Environmental Resources, North Carolina State University, Raleigh, North Carolina 9 North Carolina Museum of Natural Sciences, Raleigh, North Carolina Correspondence Elizabeth Heppenheimer, Department of Ecology & Evolutionary Biology, Princeton University, Princeton, NJ. Email: [email protected] Funding information National Science Foundation, Grant/Award Number: 1523859 Abstract Range expansion is a widespread biological process, with well‐described theoretical expectations associated with the colonization of novel ranges. However, compara‐ tively few empirical studies address the genomic outcomes accompanying the ge‐ nome‐wide consequences associated with the range expansion process, particularly in recent or ongoing expansions. Here, we assess two recent and distinct eastward expansion fronts of a highly mobile carnivore, the coyote (Canis latrans), to investi‐ gate patterns of genomic diversity and identify variants that may have been under selection during range expansion. Using a restriction‐associated DNA sequencing (RADseq), we genotyped 394 coyotes at 22,935 SNPs and found that overall popula‐ tion structure corresponded to their 19th century historical range and two distinct populations that expanded during the 20th century. Counter to theoretical expecta‐ tions for populations to bottleneck during range expansions, we observed minimal evidence for decreased genomic diversity across coyotes sampled along either ex‐ pansion front, which is likely due to hybridization with other Canis species. Furthermore, we identified 12 SNPs, located either within genes or putative regula‐ tory regions, that were consistently associated with range expansion. Of these 12 genes, three (CACNA1C, ALK, and EPHA6) have putative functions related to disper‐ sal, including habituation to novel environments and spatial learning, consistent with the expectations for traits under selection during range expansion. Although coyote colonization of eastern North America is well‐publicized, this study provides novel insights by identifying genes associated with dispersal capabilities in coyotes on the two eastern expansion fronts. KEYWORDS colonization, dispersal behavior, outlier SNPs, RADseq, range expansion

Transcript of High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough,...

Page 1: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

Ecology and Evolution 20181ndash15 emsp|emsp1wwwecolevolorg

Received28June2018emsp |emsp Revised10October2018emsp |emsp Accepted15October2018DOI101002ece34688

O R I G I N A L R E S E A R C H

High genomic diversity and candidate genes under selection associated with range expansion in eastern coyote (Canis latrans) populations

Elizabeth Heppenheimer1 emsp|emspKristin E Brzeski12emsp|emspJoseph W Hinton3 emsp|emsp Brent R Patterson45emsp|emspLinda Y Rutledge15 emsp|emspAlexandra L DeCandia1 emsp|emsp Tyler Wheeldon45emsp|emspSteven R Fain6emsp|emspPaul A Hohenlohe7emsp|emspRoland Kays89emsp|emsp Bradley N White5emsp|emspMichael J Chamberlain3emsp|emspBridgett M vonHoldt1

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicensewhichpermitsusedistributionandreproductioninanymediumprovidedtheoriginalworkisproperlycitedcopy2018TheAuthorsEcology and EvolutionpublishedbyJohnWileyampSonsLtd

1DepartmentofEcologyampEvolutionaryBiologyPrincetonUniversityPrincetonNewJersey2SchoolofForestResourcesandEnvironmentalScienceMichiganTechnologicalUniversityHoughtonMichigan3WarnellSchoolofForestryandNaturalResourcesUniversityofGeorgiaAthensGeorgia4OntarioMinistryofNaturalResourcesandForestryPeterboroughOntarioCanada5TrentUniversityPeterboroughOntario6USFWSNationalForensicsLaboratoryAshlandOregon7DepartmentofBiologicalSciencesInstituteforBioinformaticsandEvolutionaryStudiesUniversityofIdahoMoscowIdaho8DepartmentofForestryandEnvironmentalResourcesNorthCarolinaStateUniversityRaleighNorthCarolina9NorthCarolinaMuseumofNaturalSciencesRaleighNorthCarolina

CorrespondenceElizabethHeppenheimerDepartmentofEcologyampEvolutionaryBiologyPrincetonUniversityPrincetonNJEmaileh7princetonedu

Funding informationNationalScienceFoundationGrantAwardNumber1523859

AbstractRangeexpansionisawidespreadbiologicalprocesswithwell‐describedtheoreticalexpectationsassociatedwiththecolonizationofnovelrangesHowevercompara‐tively fewempirical studiesaddress thegenomicoutcomesaccompanying thege‐nome‐wideconsequencesassociatedwiththerangeexpansionprocessparticularlyinrecentorongoingexpansionsHereweassesstworecentanddistincteastwardexpansionfrontsofahighlymobilecarnivorethecoyote(Canis latrans)toinvesti‐gatepatternsofgenomicdiversityandidentifyvariantsthatmayhavebeenunderselection during range expansion Using a restriction‐associatedDNA sequencing(RADseq)wegenotyped394coyotesat22935SNPsandfoundthatoverallpopula‐tionstructurecorrespondedtotheir19thcenturyhistoricalrangeandtwodistinctpopulationsthatexpandedduringthe20thcenturyCountertotheoreticalexpecta‐tionsforpopulationstobottleneckduringrangeexpansionsweobservedminimalevidencefordecreasedgenomicdiversityacrosscoyotessampledalongeitherex‐pansion front which is likely due to hybridization with other Canis speciesFurthermoreweidentified12SNPslocatedeitherwithingenesorputativeregula‐toryregions thatwereconsistentlyassociatedwithrangeexpansionOfthese12genesthree(CACNA1CALKandEPHA6)haveputativefunctionsrelatedtodisper‐salincludinghabituationtonovelenvironmentsandspatiallearningconsistentwiththeexpectationsfortraitsunderselectionduringrangeexpansionAlthoughcoyotecolonizationofeasternNorthAmericaiswell‐publicizedthisstudyprovidesnovelinsightsbyidentifyinggenesassociatedwithdispersalcapabilitiesincoyotesonthetwoeasternexpansionfronts

K E Y W O R D S

colonizationdispersalbehavioroutlierSNPsRADseqrangeexpansion

2emsp |emsp emspensp HEPPENHEIMER Et al

1emsp |emspINTRODUC TION

Range expansions are a ubiquitous aspect of natural history(ExcoffierFollampPetit2009)Recentorongoingrangeexpansionshavebeendocumented inavarietyofdiverse taxa includingbut‐terflies(BraschlerampHill2007Hilletal2001PatemanHillRoyFoxampThomas2012)mammals(Balestrierietal2010TaulmanampRobbins1996)andbirds(Livezey2009Swaegersetal2015)aswellasnumerousplantspecies(ArianiMierampyTeranJampGeptsP2017ColauttiampBarrett2013VossEcksteinampDurka2012)Yetdespitethewidespreadprevalenceofsubstantialrangeexpan‐sionscomparativelyfewempiricalstudieshaveexploredthegeneticor genomic consequences of recent or ongoing expansions withsomeexceptions(HagenKopatzAspiKojolaampEiken2015Noreacutenetal2015Heppenheimeretal2018)

Broadly range expansion is expected to result in reduced ge‐nome‐widediversityrelativetothecorerangeasaconsequenceofsmallpopulation sizesand serial founderevents (Mayr1954NeiMaruyama amp Chakraborty 1975) Strong population structure isalso expected along the expansion axis with recently expandedpopulationsoftenrepresentingdifferentiatedgeneticclustersfromthoseinthecorerange(IbrahimandNichols1996)Thoughdemo‐graphicfactorssuchasfecundityandpopulationdensityinfluencepopulation structure and genome‐wide diversity of recently ex‐pandedpopulations(Hagenetal2015)naturalselectionoftraits(iereproductiondispersal)associatedwithrangeexpansionsmayalso play an important role Theoretically traits facilitating rangeexpansionshouldexperiencedifferentialselectionpressuresalongtheaxisofexpansion(TravisampDytham2002Phillipsetal2008BurtonPhillipsampTravis2010)Forinstancegenesassociatedwithexploratorybehavioranddispersalabilitiesarepredictedtobeben‐eficialatthefrontoftheexpansionaxisgivensuchtraitsdirectlyfa‐cilitatemovementintoandsubsequentcolonizationofanewhabitat(Burtonetal2010HughesDythamampHill2007PhillipsBrownTravisampShine2008TravisampDytham2002)Reproductivetraitsarealsopredictedtobeunderselectionasreducedcompetitionandsmallerpopulationsizesatthefrontoftheexpansionmayfavorin‐creasedreproductiveeffort(Burtonetal2010)

Althoughpredictionsforadaptiveevolutionduringrangeexpan‐sionarewelldescribedintheoryitisachallengeinpracticetoiden‐tifylociunderselectionattherangeperipheryforseveralreasonsFirstastochasticphenomenonknownasldquoallelesurfingrdquoaconse‐quence of serial founder events and drift at the expansion frontmaydriveevendeleteriousallelestohighfrequenciesalongtheex‐pansionaxisThisprocesshasastrongtheoreticalbasis(EdmondsLillieampCavalli‐Sforza2004KlopfsteinCurratampExcoffier2006)andbeensuggestedinempiricalstudiesforarangeoftaxa(HoferRayWegmannampExcoffier2009Gralkaetal2016Streicheretal2016)Therefore identifyinggenomicvariantswithsubstantialchanges in frequency along the expansion axis alone is not a suf‐ficientevidenceofrecentselectionAdditionallyvariationinallelefrequencymaybedrivenbyenvironmentalfactors(egnovelhab‐itatsandfoodresources)thatoccurintheexpandingrangebutare

independentof traits (egdispersal and reproductivecapabilities)thatdirectlyfacilitaterangeexpansion

Whiletheeffectsofallelesurfingandenvironmentalfactorscannotbecompletelyaccountedforwhenstudyingrangeexpan‐sionreplicateexpansionfrontsacrossdistinctenvironmentscanhelpdisentangletherelativeimpactoftheseforces(Swaegersetal2015WhitePerkinsHeckelampSearle2013)Asallelesurf‐ingisastochasticprocessitislesslikelythatthesamegenomicvariant would undergo a frequency shift in the same directionrelativetothehistoricalrangealongmultipleindependentexpan‐sionaxesSimilarlywhenspeciestraversedistinctenvironmentsincreasesordecreasesinfrequencyatthesamelociarelesslikelytobedrivenbylocaladaptationThereforegenomicvariantsthatundergo similar frequency shifts across multiple independentaxesofexpansionarereasonablecandidatesforrangeexpansiongenes

Coyotes (Canis latrans) provide a tractable system to addressquestions related to range expansion genomics Confined to thewesternandcentralregionsofNorthAmericapriorto1900(Nowak19792002 YoungampJackson1951)hereafter referredtoas thecoyotehistorical range coyoteshave substantially expanded theirgeographic rangeover the last century tooccupyeverycontinen‐talUSstateandCanadianprovince(HodyampKays2018)Herewefocus on the eastward expansion across the midwestern US andsoutheastern Canada culminating along the eastern seaboardThisexpansionbegan in theearly20thcenturyand followed twobroadexpansionroutesacrossdistinctenvironmentsInthenorth‐eastcoyotesexpandedacrosstheGreatLakesregionoftheUnitedStatesandCanadaintoNewEnglandNewYorkandPennsylvaniaThesoutheasternexpansionoccurredataslowerrateandfollowedanapproximatetrajectorythroughLouisianaAlabamaandGeorgiawith initial reports of coyotes in the Carolinas as recently as the1980s(DeBowWebsterampSumner1998)Thoughfine‐scalevari‐ation inexpansion routeshasbeendocumented for thenortheast(KaysCurtisampKirchman2010Wheeldonet al 2010) a recentgeneticsurvey(Heppenheimeretal2018)supportstwogeneticallydistincteasterncoyotepopulationsacrosstheeasternseaboardthatcorrespondtothesebroadlydescribednortheasternandsoutheast‐ern expansion routes suggesting that fine‐scale expansion routeshavelikelyconverged

In addition to geographic isolation each expansion front rep‐resents distinct ecoregions in North America For example thenortheasternexpansionfrontisprimarilynorthernforestsandeast‐erntemperateforestswhichisfurtherdividedprimarilyintomixedwoodshieldAtlanticHighlandsandmixedwoodplains(OmernikampGriffith2014)Incontrastthesoutheasternexpansionfrontisal‐mostentirelyeasterntemperateforestsbuttransitionstotropicalwetforestandgreatplainsdesignationsalongtheGulfofMexico(OmernikampGriffith2014)FurthermoreeachexpansionfrontalsodiffersinthepresenceandabundanceofcloselyrelatedCanisspe‐ciesGenerallyunderarangeexpansionscenariohybridizationbe‐tweencloselyrelatedandpreviouslyisolatedspeciesmayoccurasaresultoflowpopulationdensityoftheexpandingspeciesalongthe

emspensp emsp | emsp3HEPPENHEIMER Et al

rangeperiphery (Seehausen2004)Assuchcoyotehybridizationwithremnantpopulationsofeastern(C lycaon)andorgraywolves(C lupus)hasbeendocumentedalongthenortheasternexpansionroute(Kaysetal2010RutledgeGarrowayLovelessampPatterson2010 vonHoldtet al 2011 vonHoldtKaysPollingerampWayne2016)aswellaswithredwolves(C rufus)inthesoutheasternex‐pansionfront(Nowak2002)InparticularredwolvesarebelievedtobeextirpatedoutsideoftheNorthCarolinarecoveryareabuthybridizationbetweenredwolvesandcoyotesiswelldocumentedwithin that area (Bohling et al 2016HintonGittlemanManenampChamberlain2018)Furtherseveralpreviousstudieshavealsoshown that eastern coyote populations have interbred with do‐mestic dogs (Adams Leonard ampWaits 2003Wilson RutledgeWheeldon Patterson amp White 2012 Wheeldon RutledgePattersonWhite ampWilson 2013 Monzotilden Kays amp Dykhuizen2014)Whilethereisevidencethatthesehybridizationeventshavebeenadaptive (vonHoldtetal2016) it is important tonotethatthefullgenome‐wideconsequencesandthegeographicextentofinterspecieshybridizationhavenotbeendocumentedthroughouttheentireeasternrange

Overallourobjectivesweretoquantifygenomicstructureanddiversity across the historical coyote range and the two recentlyexpandedeasterncoyotepopulationsWethenidentifyoutlierlocithatmayhavebeenunderselectioninbothpopulationsasaresultofrangeexpansionWepredictthatpopulationstructurewillcorre‐spond to theknowndemographichistoryofNorthAmericancoy‐otesthatisahistoricalrangepopulationandtwodistinctrecentlyexpandedgroups In accordancewith theoretical assumptionsweexpect reduced genomic diversity in the two recently expandedeastern populations relative to the historical range However hy‐bridizationwith otherCanis speciesmay result in deviations fromthis expectation Finally we expect genomic variants that under‐wentfrequencyshiftsinthesamedirectioninbothgroupstohaveputativefunctionsrelatedtorangeexpansionsuchasdispersalandreproduction

2emsp |emspMETHODS

21emsp|emspSample collection

Weobtainedcoyotebloodand tissue (eg liver kidney tongue)fromstatemanagementprograms(PrincetonIACUC1961A‐13)governmentorganizationarchives(egFloridaFishandWildlifeUS Department of Agriculture Ontario Ministry of NaturalResources and Forestry) or museum archives (New York StateMuseum Oklahoma Museum of Natural History) In all casesstateorprovinceoforiginwasdocumented(Figure1)andinmanycasessexapproximateageandfine‐scalegeographicdatawerealso known Samples were collected between 1998 and 2017withthemajoritycollectedwithinthelast10years(2008ndash2017)Sampleswithunknowncollectiondateswereeitherknownoras‐sumedtofallwithintheapproximatetimeperiod(Supportingin‐formationTableS1)

For downstream analyses we considered samples collectedfrom AZ CA IDMNMO NE NM NV OK SK TXWA andWYtobepartofthehistoricalrange(iepre‐1900Figure1)asdescribed by Hody and Kays (2018) Additionally samples col‐lectedfromMENBNJONandPAwereconsideredpartofthenortheastexpansionandsamplescollectedfromALFLGAKYLANCSCTNandVAwereconsideredpartofthesoutheasternexpansion

22emsp|emspSampling and DNA extraction

HighmolecularweightgenomicDNAwasextractedfromallsampleswitheithertheQiagenDNeasyBloodandTissueKitortheBioSprint96DNABloodKitinconjunctionwithaThermoScientificKingFisherFlexPurificationplatforminbothcasesfollowinginstructionspro‐videdbythemanufacturerWequantifiedDNAconcentrationwitheitherPicoGreenorQubit20 fluorometryand standardized sam‐ples to5ngμlOnlyhigh‐qualityDNAsamples asdeterminedbythepresenceofahighmolecularweightbandona1agarosegelwereretainedforsequencing

23emsp|emspRADsequencing and bioinformatics processing

Wepreparedgenomiclibrariesfollowingamodifiedrestriction‐as‐sociatedDNAsequencing(RADseq)protocoldescribedbyAlietal(2016)BrieflysamplesweredigestedwithsbfIandaunique8bpbarcoded biotinylated adapter was ligated to the resulting frag‐mentsSampleswerethenpooled(96ndash153samplespool)andran‐domlyshearedto400bpinaCovarisLE220FollowingshearingweusedaDynabeadsM‐280streptavidinbeadbindingassaytoenrichforadapter‐ligatedfragmentsFinalsequencinglibrarieswerethenpreparedusingeithertheNEBNextUltraDNALibraryPrepKitortheNEBNextUltraIIDNALibraryPrepKitSizeselectionwasmadefor300ndash400bpinsertwithAgencourtAMPureXPmagneticbeadswhichwerealsousedforlibrarypurificationLibrarieswerestand‐ardized to 10nM and sequenced (2X150nt) on two lanes on theIlluminaHiSeq2500

As thisRADseqprotocol isunique in that thebarcodemaybeoneithertheforwardorreversereaddataprocessingwasrequiredprior to variant calling Accordingly forward and reverse raw se‐quencing readswereprocessedsuch thatany readcontaining theremnantsbfIcutsiteandoneofthepossiblebarcodeswerealignedinasinglefilewhilethematchingreadpairsthatlackedthecutsitewere aligned in a separate file and all remaining reads were dis‐cardedThiswasaccomplishedusingacustomPerlscript(flip_trim_sbfI_170601plseeSupportinginformation)

AdditionaldataprocessingwasthenconductedinSTACKSv142 (CatchenHohenloheBasshamAmoresampCresko2013)First readsweredemultiplexedusingprocess_radtags allowinga 2bpmismatch for barcode rescue and discarding readswitheitheruncalledbasesora low‐qualityscore (lt10)withinaslid‐ing window of 015 Next PCR duplicates were removedwiththe paired‐end sequencing filtering option in clone_filter We

4emsp |emsp emspensp HEPPENHEIMER Et al

excluded all sampleswith low read count (lt500000) after re‐movalofPCRduplicates fromfurtheranalysisRemainingsam‐pleswerethenmappedtothedoggenomeCanFam31assembly(Lindblad‐Tohetal2005)inStampyv1021(LunterampGoodson2011)Toreducesbiasesassociatedwithincorrectlymappedloci(egparalogs)weadditionallyfilteredmappedreadsforamin‐imumMAPQof96andconvertedtobamformat inSamtoolsv0118(Lietal2009)

We completed SNP calling in STACKS following the rec‐ommended pipeline for referencemapped data (iepstacks rarr cstacks rarr sstacks rarr populations Catchenetal2013)Inpstackswe requiredaminimumdepthofcoverageof three to reportastack (‐m3)Cstacks and sstackswere run as recommendedbyCatchenetal (2013)ThepopulationsmodulewasruntwicetooptimizefinalsampleselectiontoreducebothmissingdataandbiasesresultingfromunevensamplingacrosslocationsFirstweallowedonlythefirstSNPperlocus(‐‐write_single_snp)tobere‐portedbutdidnotapplyanymissingdata thresholdsWethenevaluatedtheper‐individualgenotypingsuccessasmeasuredbymissingnessperindividualinPlinkv190b3i(Purcelletal2007)andremovedindividualswithahighlevelofmissingdata(gt80missing)Wealsoremovedsamplesfromlocationswheren = 1 Inoursecondrunofpopulationsweincludedonlythisreducedsetofsamplesrequiredthatreportedlocibegenotypedin90ofindividuals(minusr=09)andagainrestrictedanalysistoonlythe

firstSNPperlocusOnlythislatterdatasetwasusedforsubse‐quentanalysesFollowingSNPcallingwefilteredforstatisticallinkagedisequilibriuminPlinkwiththeargument‐‐indep‐pairwise 50 5 05

All SNPs were annotated as genic (intron or exon) within apromoter(iewithin2Kboftranscriptionstartsitefollowingvon‐HoldtHeppenheimerPetrenkoCroonquistandRutledge(2017))orintergenicusinganin‐housepythonscript(chr_sitepySeesup‐porting information)All intergenic SNPswere compiled in a sec‐ondgenotypedatasetandfilteredforHardyndashWeinbergEquilibrium(HWE) in Plinkwith the argumentmdashhwe 0001 These intergenicHWE‐filtered genotypes were presumed neutral in downstreamanalyses (hereafterputativelyneutral loci)Additionallywecom‐piled a third dataset of putatively functional loci consistingof allSNPsannotatedasgenicorwithin2Kbofatranscriptionstartsite(hereaftergenicloci)

24emsp|emspPopulation structure analysis

Tovisualize clustering inourdata and identify strongoutliersweconductedaPrincipalComponentAnalysisusingourfullSNPdata‐set with flashPCA (Abraham amp Inouye 2014) We identified onestrongoutlieroriginating fromOntariowhichmaybeamisidenti‐fiedeasternorgraywolf (C lycaon or C lupus)This individualwasremovedfromfurtheranalyses

F I G U R E 1 emspMapofcoyotehistoricalrange1950srange2000srangeandsamplesizeperlocationRangesareapproximateandmodifiedfromHodyandKays(2018)

AL GA

FL

SC

LA

1950s2000s

29

13

10

26

12

4

4

11

20

2

5

12

6

4

28

16 23

13

27

25

373

144

Historical range (pre-1900)

14

26

2

4

NCVVAVV

PAPP NJNJNY

MEMENB

ON

TTXX

AZAZ

MNMN

CCAA

WYWWYY

IDD

WWWAAWWW

NVNV

SKSSK

OKOOK

NENEEMOMMOO

KKKKKKKYKTNN

NMNM

emspensp emsp | emsp5HEPPENHEIMER Et al

Following the removal of strong outliers (eg putatively mis‐identified wolves) we determined the most likely number of ge‐nomicclustersrepresentedbythedatabyconductingananalysisofpopulationstructureinADMIXTUREv13(AlexanderNovembreampLange2009)withthecross‐validationflagWeevaluatedK=1ndash10withtheKvaluewiththelowestcross‐validation(cv)scoreindicativeof thebest fitKADMIXTURE is similar inprinciple to theclassicBayesian software STRUCTURE (Pritchard Stephens ampDonnelly2000)butusesamaximumlikelihoodframeworkandismorecom‐putationallyefficientforSNPdata

25emsp|emspGenomic diversity

Standardmetricsofgenomicdiversityforeachsamplinglocationin‐cludingprivateallelecountsaswellasobserved(Ho)andexpected(He) heterozygositywerecalculatedinSTACKSacrossall lociTodeter‐minewhethertrendsofheterozygositydifferedbasedonwhichpartofthegenomewassurveyedwerecalculatedbothHoandHeacrossputativelyneutrallociandgeniclociAllelicrichness(Ar)andprivateallelicrichness(Apr)werecalculatedusingararefactionapproachim‐plementedinADZEv10(SzpiechJakobssonampRosenberg2008)wherethemaximumstandardizedsamplesizewassettothesmallestnforthesamplesconsidered(ie176whencomparingthehistoricalrangenortheastexpansionandsoutheastexpansion)

PairwiseFSTvaluesbetweenall sampling locationsoverall lociwerecalculatedinSTACKSandwetestedforisolationbydistance(IBD)within thecoyotehistorical rangeaswellaswithineachre‐cently expanded eastern populationwith a series ofMantel testsimplementedinade4v17‐11(DrayampDufour2007)inRv33(RCoreTeam2013)PairwiseFSTwerelinearizedfollowingRousset(1997)geographic distanceswere calculated as the shortest straight‐linedistancebetweensamplinglocationsandsignificancewasassessedfrom9999permutations

26emsp|emspIdentification of loci associated with range expansion

Toidentifylociascandidatesforselectionduringrangeexpansionwerestrictedanalysestoindividualswithhighclusterassignmentsas identified intheADMIXTUREanalysis (Qge08)Additionallytopreventbiasesresultingfromlong‐distancedispersersweremovedthreeindividualsthathadhighclusterassignmentstodifferentpop‐ulationsfromwhichtheyweresampled(egsampledfromLouisianabut clustered with the northeast group) Though restricting theanalysestoindividualswithhighclusterassignmentsmayinflatethegeneticdistinctionbetweenthehistoricalrangeandtherecentlyex‐pandedcoyotepopulationswebelievetheinferredhistoricalrangepopulationtobethebestavailablerepresentationofpre‐rangeex‐pansionallelefrequenciesasthereis likelyongoingcontemporarygeneflowbetweencoyotesinthehistoricalrangeandbothrecentlyexpandedpopulations

AsFSToutlier‐typeapproachestoidentifylociunderselectionarepronetohighratesoffalsepositivesespeciallyamongpopulations

with complex demographic histories we used two distinct ap‐proachestoidentifylociputativelyunderselectionFirstaBayesianframework to detect outlier loci was implemented in BAYENV2(CoopWitonskyRienzoampPritchard2010GuumlntherampCoop2013)Thismethodaccountsforevolutionarynonindependencebetweenpopulationsbyfirstcalculatingcovarianceinallelefrequenciesataset of putatively neutral loci Candidate functional SNPs are thenevaluatedoneata timeunderamodel thatassumesa linear rela‐tionship between an environmental variable and allele frequencycomparedtoamodelgivenbytheneutralcovariancematrixandacorrespondingBayesfactoriscalculatedThismethodhasbeensug‐gested to outperformotherFSToutlier‐likemethods (eg FDIST2BayeScan) in the case of range expansion (Lotterhos ampWhitlock2014) In theBAYENV2analysis theenvironmentalvariableof in‐terestwasthelineardistancefromthecoyotehistoricalrange(egWhiteetal2013)Toavoidbiasesinducedbyallelefrequenciesatanyonesampling locationwithin thehistorical rangeall statesorprovinceswithinthehistoricalcoyoterangeweretreatedasasinglesampling locationDistances for sampling locationsoutsideof thehistoricalrangewerecalculatedastheshorteststraight‐linedistancefrom theapproximatemidpointof the sampled regionswithin thehistoricalrangeThenorthernandsouthernexpansionfrontswereanalyzedseparately Inbothcasestheputativelyneutral lociusedtogeneratethecontrolcovariancematrixwere intergenicSNPs inHWEfilteredfurthertoremoveanymonomorphiclocibetweenthepopulationscomparedSimilarlythecandidateSNPswereallSNPsannotatedasgenic(intronorexon)orwithin2Kbofatranscriptionstartsiteagainfilteredtoremovemonomorphiclocibetweenpop‐ulationsGenotype files forbothSNPdatasetswereconverted toBAYENV2formatinPGDSpiderv2113(LischerampExcoffier2012)For each expansion front (historical to northeast amp historical tosoutheast)lociwererankedbyBayesfactorandthetop3ofSNPswereretainedforfurtheranalysisLociwereconsideredcandidategenesunderselectionduringrangeexpansionifthesameSNPoc‐curredonbothlists

Secondweusedaprincipalcomponent‐basedapproachimple‐mentedwith theRpackagePCadapt (LuuBazinampBlum2017)This method first performs a centered scaled principal compo‐nentanalysisongenome‐wideSNPsandthenidentifiessignificantoutlierswith respect topopulation structuregivenby the firstK principal components Specifically PCadapt identifies outliersbasedon theMahalanobisdistancewhichdescribesmultidimen‐sionaldistanceofapointfromthemeanSimulationsindicatethatPCadaptislesspronetotypeIIerrorthanalternativemethods(egBayeScan)andPCadaptisexpectedtoperformwellunderavarietyofcomplexdemographicscenariosincludingrangeexpansion(Luuetal2017)

InthePCadaptanalysisthenortheasternandsoutheasternex‐pansionfrontswereanalyzedseparatelyInbothcasesweidenti‐fiedoutlierswithrespecttounderlyingpopulationstructuregivenby PC1We chose to retain only PC1 rather than selecting theoptimalnumberofPCsbasedonconventionalmethods(egscreeplot)asPC1primarilycapturedthemajoraxisofrangeexpansion

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 2: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

2emsp |emsp emspensp HEPPENHEIMER Et al

1emsp |emspINTRODUC TION

Range expansions are a ubiquitous aspect of natural history(ExcoffierFollampPetit2009)Recentorongoingrangeexpansionshavebeendocumented inavarietyofdiverse taxa includingbut‐terflies(BraschlerampHill2007Hilletal2001PatemanHillRoyFoxampThomas2012)mammals(Balestrierietal2010TaulmanampRobbins1996)andbirds(Livezey2009Swaegersetal2015)aswellasnumerousplantspecies(ArianiMierampyTeranJampGeptsP2017ColauttiampBarrett2013VossEcksteinampDurka2012)Yetdespitethewidespreadprevalenceofsubstantialrangeexpan‐sionscomparativelyfewempiricalstudieshaveexploredthegeneticor genomic consequences of recent or ongoing expansions withsomeexceptions(HagenKopatzAspiKojolaampEiken2015Noreacutenetal2015Heppenheimeretal2018)

Broadly range expansion is expected to result in reduced ge‐nome‐widediversityrelativetothecorerangeasaconsequenceofsmallpopulation sizesand serial founderevents (Mayr1954NeiMaruyama amp Chakraborty 1975) Strong population structure isalso expected along the expansion axis with recently expandedpopulationsoftenrepresentingdifferentiatedgeneticclustersfromthoseinthecorerange(IbrahimandNichols1996)Thoughdemo‐graphicfactorssuchasfecundityandpopulationdensityinfluencepopulation structure and genome‐wide diversity of recently ex‐pandedpopulations(Hagenetal2015)naturalselectionoftraits(iereproductiondispersal)associatedwithrangeexpansionsmayalso play an important role Theoretically traits facilitating rangeexpansionshouldexperiencedifferentialselectionpressuresalongtheaxisofexpansion(TravisampDytham2002Phillipsetal2008BurtonPhillipsampTravis2010)Forinstancegenesassociatedwithexploratorybehavioranddispersalabilitiesarepredictedtobeben‐eficialatthefrontoftheexpansionaxisgivensuchtraitsdirectlyfa‐cilitatemovementintoandsubsequentcolonizationofanewhabitat(Burtonetal2010HughesDythamampHill2007PhillipsBrownTravisampShine2008TravisampDytham2002)Reproductivetraitsarealsopredictedtobeunderselectionasreducedcompetitionandsmallerpopulationsizesatthefrontoftheexpansionmayfavorin‐creasedreproductiveeffort(Burtonetal2010)

Althoughpredictionsforadaptiveevolutionduringrangeexpan‐sionarewelldescribedintheoryitisachallengeinpracticetoiden‐tifylociunderselectionattherangeperipheryforseveralreasonsFirstastochasticphenomenonknownasldquoallelesurfingrdquoaconse‐quence of serial founder events and drift at the expansion frontmaydriveevendeleteriousallelestohighfrequenciesalongtheex‐pansionaxisThisprocesshasastrongtheoreticalbasis(EdmondsLillieampCavalli‐Sforza2004KlopfsteinCurratampExcoffier2006)andbeensuggestedinempiricalstudiesforarangeoftaxa(HoferRayWegmannampExcoffier2009Gralkaetal2016Streicheretal2016)Therefore identifyinggenomicvariantswithsubstantialchanges in frequency along the expansion axis alone is not a suf‐ficientevidenceofrecentselectionAdditionallyvariationinallelefrequencymaybedrivenbyenvironmentalfactors(egnovelhab‐itatsandfoodresources)thatoccurintheexpandingrangebutare

independentof traits (egdispersal and reproductivecapabilities)thatdirectlyfacilitaterangeexpansion

Whiletheeffectsofallelesurfingandenvironmentalfactorscannotbecompletelyaccountedforwhenstudyingrangeexpan‐sionreplicateexpansionfrontsacrossdistinctenvironmentscanhelpdisentangletherelativeimpactoftheseforces(Swaegersetal2015WhitePerkinsHeckelampSearle2013)Asallelesurf‐ingisastochasticprocessitislesslikelythatthesamegenomicvariant would undergo a frequency shift in the same directionrelativetothehistoricalrangealongmultipleindependentexpan‐sionaxesSimilarlywhenspeciestraversedistinctenvironmentsincreasesordecreasesinfrequencyatthesamelociarelesslikelytobedrivenbylocaladaptationThereforegenomicvariantsthatundergo similar frequency shifts across multiple independentaxesofexpansionarereasonablecandidatesforrangeexpansiongenes

Coyotes (Canis latrans) provide a tractable system to addressquestions related to range expansion genomics Confined to thewesternandcentralregionsofNorthAmericapriorto1900(Nowak19792002 YoungampJackson1951)hereafter referredtoas thecoyotehistorical range coyoteshave substantially expanded theirgeographic rangeover the last century tooccupyeverycontinen‐talUSstateandCanadianprovince(HodyampKays2018)Herewefocus on the eastward expansion across the midwestern US andsoutheastern Canada culminating along the eastern seaboardThisexpansionbegan in theearly20thcenturyand followed twobroadexpansionroutesacrossdistinctenvironmentsInthenorth‐eastcoyotesexpandedacrosstheGreatLakesregionoftheUnitedStatesandCanadaintoNewEnglandNewYorkandPennsylvaniaThesoutheasternexpansionoccurredataslowerrateandfollowedanapproximatetrajectorythroughLouisianaAlabamaandGeorgiawith initial reports of coyotes in the Carolinas as recently as the1980s(DeBowWebsterampSumner1998)Thoughfine‐scalevari‐ation inexpansion routeshasbeendocumented for thenortheast(KaysCurtisampKirchman2010Wheeldonet al 2010) a recentgeneticsurvey(Heppenheimeretal2018)supportstwogeneticallydistincteasterncoyotepopulationsacrosstheeasternseaboardthatcorrespondtothesebroadlydescribednortheasternandsoutheast‐ern expansion routes suggesting that fine‐scale expansion routeshavelikelyconverged

In addition to geographic isolation each expansion front rep‐resents distinct ecoregions in North America For example thenortheasternexpansionfrontisprimarilynorthernforestsandeast‐erntemperateforestswhichisfurtherdividedprimarilyintomixedwoodshieldAtlanticHighlandsandmixedwoodplains(OmernikampGriffith2014)Incontrastthesoutheasternexpansionfrontisal‐mostentirelyeasterntemperateforestsbuttransitionstotropicalwetforestandgreatplainsdesignationsalongtheGulfofMexico(OmernikampGriffith2014)FurthermoreeachexpansionfrontalsodiffersinthepresenceandabundanceofcloselyrelatedCanisspe‐ciesGenerallyunderarangeexpansionscenariohybridizationbe‐tweencloselyrelatedandpreviouslyisolatedspeciesmayoccurasaresultoflowpopulationdensityoftheexpandingspeciesalongthe

emspensp emsp | emsp3HEPPENHEIMER Et al

rangeperiphery (Seehausen2004)Assuchcoyotehybridizationwithremnantpopulationsofeastern(C lycaon)andorgraywolves(C lupus)hasbeendocumentedalongthenortheasternexpansionroute(Kaysetal2010RutledgeGarrowayLovelessampPatterson2010 vonHoldtet al 2011 vonHoldtKaysPollingerampWayne2016)aswellaswithredwolves(C rufus)inthesoutheasternex‐pansionfront(Nowak2002)InparticularredwolvesarebelievedtobeextirpatedoutsideoftheNorthCarolinarecoveryareabuthybridizationbetweenredwolvesandcoyotesiswelldocumentedwithin that area (Bohling et al 2016HintonGittlemanManenampChamberlain2018)Furtherseveralpreviousstudieshavealsoshown that eastern coyote populations have interbred with do‐mestic dogs (Adams Leonard ampWaits 2003Wilson RutledgeWheeldon Patterson amp White 2012 Wheeldon RutledgePattersonWhite ampWilson 2013 Monzotilden Kays amp Dykhuizen2014)Whilethereisevidencethatthesehybridizationeventshavebeenadaptive (vonHoldtetal2016) it is important tonotethatthefullgenome‐wideconsequencesandthegeographicextentofinterspecieshybridizationhavenotbeendocumentedthroughouttheentireeasternrange

Overallourobjectivesweretoquantifygenomicstructureanddiversity across the historical coyote range and the two recentlyexpandedeasterncoyotepopulationsWethenidentifyoutlierlocithatmayhavebeenunderselectioninbothpopulationsasaresultofrangeexpansionWepredictthatpopulationstructurewillcorre‐spond to theknowndemographichistoryofNorthAmericancoy‐otesthatisahistoricalrangepopulationandtwodistinctrecentlyexpandedgroups In accordancewith theoretical assumptionsweexpect reduced genomic diversity in the two recently expandedeastern populations relative to the historical range However hy‐bridizationwith otherCanis speciesmay result in deviations fromthis expectation Finally we expect genomic variants that under‐wentfrequencyshiftsinthesamedirectioninbothgroupstohaveputativefunctionsrelatedtorangeexpansionsuchasdispersalandreproduction

2emsp |emspMETHODS

21emsp|emspSample collection

Weobtainedcoyotebloodand tissue (eg liver kidney tongue)fromstatemanagementprograms(PrincetonIACUC1961A‐13)governmentorganizationarchives(egFloridaFishandWildlifeUS Department of Agriculture Ontario Ministry of NaturalResources and Forestry) or museum archives (New York StateMuseum Oklahoma Museum of Natural History) In all casesstateorprovinceoforiginwasdocumented(Figure1)andinmanycasessexapproximateageandfine‐scalegeographicdatawerealso known Samples were collected between 1998 and 2017withthemajoritycollectedwithinthelast10years(2008ndash2017)Sampleswithunknowncollectiondateswereeitherknownoras‐sumedtofallwithintheapproximatetimeperiod(Supportingin‐formationTableS1)

For downstream analyses we considered samples collectedfrom AZ CA IDMNMO NE NM NV OK SK TXWA andWYtobepartofthehistoricalrange(iepre‐1900Figure1)asdescribed by Hody and Kays (2018) Additionally samples col‐lectedfromMENBNJONandPAwereconsideredpartofthenortheastexpansionandsamplescollectedfromALFLGAKYLANCSCTNandVAwereconsideredpartofthesoutheasternexpansion

22emsp|emspSampling and DNA extraction

HighmolecularweightgenomicDNAwasextractedfromallsampleswitheithertheQiagenDNeasyBloodandTissueKitortheBioSprint96DNABloodKitinconjunctionwithaThermoScientificKingFisherFlexPurificationplatforminbothcasesfollowinginstructionspro‐videdbythemanufacturerWequantifiedDNAconcentrationwitheitherPicoGreenorQubit20 fluorometryand standardized sam‐ples to5ngμlOnlyhigh‐qualityDNAsamples asdeterminedbythepresenceofahighmolecularweightbandona1agarosegelwereretainedforsequencing

23emsp|emspRADsequencing and bioinformatics processing

Wepreparedgenomiclibrariesfollowingamodifiedrestriction‐as‐sociatedDNAsequencing(RADseq)protocoldescribedbyAlietal(2016)BrieflysamplesweredigestedwithsbfIandaunique8bpbarcoded biotinylated adapter was ligated to the resulting frag‐mentsSampleswerethenpooled(96ndash153samplespool)andran‐domlyshearedto400bpinaCovarisLE220FollowingshearingweusedaDynabeadsM‐280streptavidinbeadbindingassaytoenrichforadapter‐ligatedfragmentsFinalsequencinglibrarieswerethenpreparedusingeithertheNEBNextUltraDNALibraryPrepKitortheNEBNextUltraIIDNALibraryPrepKitSizeselectionwasmadefor300ndash400bpinsertwithAgencourtAMPureXPmagneticbeadswhichwerealsousedforlibrarypurificationLibrarieswerestand‐ardized to 10nM and sequenced (2X150nt) on two lanes on theIlluminaHiSeq2500

As thisRADseqprotocol isunique in that thebarcodemaybeoneithertheforwardorreversereaddataprocessingwasrequiredprior to variant calling Accordingly forward and reverse raw se‐quencing readswereprocessedsuch thatany readcontaining theremnantsbfIcutsiteandoneofthepossiblebarcodeswerealignedinasinglefilewhilethematchingreadpairsthatlackedthecutsitewere aligned in a separate file and all remaining reads were dis‐cardedThiswasaccomplishedusingacustomPerlscript(flip_trim_sbfI_170601plseeSupportinginformation)

AdditionaldataprocessingwasthenconductedinSTACKSv142 (CatchenHohenloheBasshamAmoresampCresko2013)First readsweredemultiplexedusingprocess_radtags allowinga 2bpmismatch for barcode rescue and discarding readswitheitheruncalledbasesora low‐qualityscore (lt10)withinaslid‐ing window of 015 Next PCR duplicates were removedwiththe paired‐end sequencing filtering option in clone_filter We

4emsp |emsp emspensp HEPPENHEIMER Et al

excluded all sampleswith low read count (lt500000) after re‐movalofPCRduplicates fromfurtheranalysisRemainingsam‐pleswerethenmappedtothedoggenomeCanFam31assembly(Lindblad‐Tohetal2005)inStampyv1021(LunterampGoodson2011)Toreducesbiasesassociatedwithincorrectlymappedloci(egparalogs)weadditionallyfilteredmappedreadsforamin‐imumMAPQof96andconvertedtobamformat inSamtoolsv0118(Lietal2009)

We completed SNP calling in STACKS following the rec‐ommended pipeline for referencemapped data (iepstacks rarr cstacks rarr sstacks rarr populations Catchenetal2013)Inpstackswe requiredaminimumdepthofcoverageof three to reportastack (‐m3)Cstacks and sstackswere run as recommendedbyCatchenetal (2013)ThepopulationsmodulewasruntwicetooptimizefinalsampleselectiontoreducebothmissingdataandbiasesresultingfromunevensamplingacrosslocationsFirstweallowedonlythefirstSNPperlocus(‐‐write_single_snp)tobere‐portedbutdidnotapplyanymissingdata thresholdsWethenevaluatedtheper‐individualgenotypingsuccessasmeasuredbymissingnessperindividualinPlinkv190b3i(Purcelletal2007)andremovedindividualswithahighlevelofmissingdata(gt80missing)Wealsoremovedsamplesfromlocationswheren = 1 Inoursecondrunofpopulationsweincludedonlythisreducedsetofsamplesrequiredthatreportedlocibegenotypedin90ofindividuals(minusr=09)andagainrestrictedanalysistoonlythe

firstSNPperlocusOnlythislatterdatasetwasusedforsubse‐quentanalysesFollowingSNPcallingwefilteredforstatisticallinkagedisequilibriuminPlinkwiththeargument‐‐indep‐pairwise 50 5 05

All SNPs were annotated as genic (intron or exon) within apromoter(iewithin2Kboftranscriptionstartsitefollowingvon‐HoldtHeppenheimerPetrenkoCroonquistandRutledge(2017))orintergenicusinganin‐housepythonscript(chr_sitepySeesup‐porting information)All intergenic SNPswere compiled in a sec‐ondgenotypedatasetandfilteredforHardyndashWeinbergEquilibrium(HWE) in Plinkwith the argumentmdashhwe 0001 These intergenicHWE‐filtered genotypes were presumed neutral in downstreamanalyses (hereafterputativelyneutral loci)Additionallywecom‐piled a third dataset of putatively functional loci consistingof allSNPsannotatedasgenicorwithin2Kbofatranscriptionstartsite(hereaftergenicloci)

24emsp|emspPopulation structure analysis

Tovisualize clustering inourdata and identify strongoutliersweconductedaPrincipalComponentAnalysisusingourfullSNPdata‐set with flashPCA (Abraham amp Inouye 2014) We identified onestrongoutlieroriginating fromOntariowhichmaybeamisidenti‐fiedeasternorgraywolf (C lycaon or C lupus)This individualwasremovedfromfurtheranalyses

F I G U R E 1 emspMapofcoyotehistoricalrange1950srange2000srangeandsamplesizeperlocationRangesareapproximateandmodifiedfromHodyandKays(2018)

AL GA

FL

SC

LA

1950s2000s

29

13

10

26

12

4

4

11

20

2

5

12

6

4

28

16 23

13

27

25

373

144

Historical range (pre-1900)

14

26

2

4

NCVVAVV

PAPP NJNJNY

MEMENB

ON

TTXX

AZAZ

MNMN

CCAA

WYWWYY

IDD

WWWAAWWW

NVNV

SKSSK

OKOOK

NENEEMOMMOO

KKKKKKKYKTNN

NMNM

emspensp emsp | emsp5HEPPENHEIMER Et al

Following the removal of strong outliers (eg putatively mis‐identified wolves) we determined the most likely number of ge‐nomicclustersrepresentedbythedatabyconductingananalysisofpopulationstructureinADMIXTUREv13(AlexanderNovembreampLange2009)withthecross‐validationflagWeevaluatedK=1ndash10withtheKvaluewiththelowestcross‐validation(cv)scoreindicativeof thebest fitKADMIXTURE is similar inprinciple to theclassicBayesian software STRUCTURE (Pritchard Stephens ampDonnelly2000)butusesamaximumlikelihoodframeworkandismorecom‐putationallyefficientforSNPdata

25emsp|emspGenomic diversity

Standardmetricsofgenomicdiversityforeachsamplinglocationin‐cludingprivateallelecountsaswellasobserved(Ho)andexpected(He) heterozygositywerecalculatedinSTACKSacrossall lociTodeter‐minewhethertrendsofheterozygositydifferedbasedonwhichpartofthegenomewassurveyedwerecalculatedbothHoandHeacrossputativelyneutrallociandgeniclociAllelicrichness(Ar)andprivateallelicrichness(Apr)werecalculatedusingararefactionapproachim‐plementedinADZEv10(SzpiechJakobssonampRosenberg2008)wherethemaximumstandardizedsamplesizewassettothesmallestnforthesamplesconsidered(ie176whencomparingthehistoricalrangenortheastexpansionandsoutheastexpansion)

PairwiseFSTvaluesbetweenall sampling locationsoverall lociwerecalculatedinSTACKSandwetestedforisolationbydistance(IBD)within thecoyotehistorical rangeaswellaswithineachre‐cently expanded eastern populationwith a series ofMantel testsimplementedinade4v17‐11(DrayampDufour2007)inRv33(RCoreTeam2013)PairwiseFSTwerelinearizedfollowingRousset(1997)geographic distanceswere calculated as the shortest straight‐linedistancebetweensamplinglocationsandsignificancewasassessedfrom9999permutations

26emsp|emspIdentification of loci associated with range expansion

Toidentifylociascandidatesforselectionduringrangeexpansionwerestrictedanalysestoindividualswithhighclusterassignmentsas identified intheADMIXTUREanalysis (Qge08)Additionallytopreventbiasesresultingfromlong‐distancedispersersweremovedthreeindividualsthathadhighclusterassignmentstodifferentpop‐ulationsfromwhichtheyweresampled(egsampledfromLouisianabut clustered with the northeast group) Though restricting theanalysestoindividualswithhighclusterassignmentsmayinflatethegeneticdistinctionbetweenthehistoricalrangeandtherecentlyex‐pandedcoyotepopulationswebelievetheinferredhistoricalrangepopulationtobethebestavailablerepresentationofpre‐rangeex‐pansionallelefrequenciesasthereis likelyongoingcontemporarygeneflowbetweencoyotesinthehistoricalrangeandbothrecentlyexpandedpopulations

AsFSToutlier‐typeapproachestoidentifylociunderselectionarepronetohighratesoffalsepositivesespeciallyamongpopulations

with complex demographic histories we used two distinct ap‐proachestoidentifylociputativelyunderselectionFirstaBayesianframework to detect outlier loci was implemented in BAYENV2(CoopWitonskyRienzoampPritchard2010GuumlntherampCoop2013)Thismethodaccountsforevolutionarynonindependencebetweenpopulationsbyfirstcalculatingcovarianceinallelefrequenciesataset of putatively neutral loci Candidate functional SNPs are thenevaluatedoneata timeunderamodel thatassumesa linear rela‐tionship between an environmental variable and allele frequencycomparedtoamodelgivenbytheneutralcovariancematrixandacorrespondingBayesfactoriscalculatedThismethodhasbeensug‐gested to outperformotherFSToutlier‐likemethods (eg FDIST2BayeScan) in the case of range expansion (Lotterhos ampWhitlock2014) In theBAYENV2analysis theenvironmentalvariableof in‐terestwasthelineardistancefromthecoyotehistoricalrange(egWhiteetal2013)Toavoidbiasesinducedbyallelefrequenciesatanyonesampling locationwithin thehistorical rangeall statesorprovinceswithinthehistoricalcoyoterangeweretreatedasasinglesampling locationDistances for sampling locationsoutsideof thehistoricalrangewerecalculatedastheshorteststraight‐linedistancefrom theapproximatemidpointof the sampled regionswithin thehistoricalrangeThenorthernandsouthernexpansionfrontswereanalyzedseparately Inbothcasestheputativelyneutral lociusedtogeneratethecontrolcovariancematrixwere intergenicSNPs inHWEfilteredfurthertoremoveanymonomorphiclocibetweenthepopulationscomparedSimilarlythecandidateSNPswereallSNPsannotatedasgenic(intronorexon)orwithin2Kbofatranscriptionstartsiteagainfilteredtoremovemonomorphiclocibetweenpop‐ulationsGenotype files forbothSNPdatasetswereconverted toBAYENV2formatinPGDSpiderv2113(LischerampExcoffier2012)For each expansion front (historical to northeast amp historical tosoutheast)lociwererankedbyBayesfactorandthetop3ofSNPswereretainedforfurtheranalysisLociwereconsideredcandidategenesunderselectionduringrangeexpansionifthesameSNPoc‐curredonbothlists

Secondweusedaprincipalcomponent‐basedapproachimple‐mentedwith theRpackagePCadapt (LuuBazinampBlum2017)This method first performs a centered scaled principal compo‐nentanalysisongenome‐wideSNPsandthenidentifiessignificantoutlierswith respect topopulation structuregivenby the firstK principal components Specifically PCadapt identifies outliersbasedon theMahalanobisdistancewhichdescribesmultidimen‐sionaldistanceofapointfromthemeanSimulationsindicatethatPCadaptislesspronetotypeIIerrorthanalternativemethods(egBayeScan)andPCadaptisexpectedtoperformwellunderavarietyofcomplexdemographicscenariosincludingrangeexpansion(Luuetal2017)

InthePCadaptanalysisthenortheasternandsoutheasternex‐pansionfrontswereanalyzedseparatelyInbothcasesweidenti‐fiedoutlierswithrespecttounderlyingpopulationstructuregivenby PC1We chose to retain only PC1 rather than selecting theoptimalnumberofPCsbasedonconventionalmethods(egscreeplot)asPC1primarilycapturedthemajoraxisofrangeexpansion

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 3: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp3HEPPENHEIMER Et al

rangeperiphery (Seehausen2004)Assuchcoyotehybridizationwithremnantpopulationsofeastern(C lycaon)andorgraywolves(C lupus)hasbeendocumentedalongthenortheasternexpansionroute(Kaysetal2010RutledgeGarrowayLovelessampPatterson2010 vonHoldtet al 2011 vonHoldtKaysPollingerampWayne2016)aswellaswithredwolves(C rufus)inthesoutheasternex‐pansionfront(Nowak2002)InparticularredwolvesarebelievedtobeextirpatedoutsideoftheNorthCarolinarecoveryareabuthybridizationbetweenredwolvesandcoyotesiswelldocumentedwithin that area (Bohling et al 2016HintonGittlemanManenampChamberlain2018)Furtherseveralpreviousstudieshavealsoshown that eastern coyote populations have interbred with do‐mestic dogs (Adams Leonard ampWaits 2003Wilson RutledgeWheeldon Patterson amp White 2012 Wheeldon RutledgePattersonWhite ampWilson 2013 Monzotilden Kays amp Dykhuizen2014)Whilethereisevidencethatthesehybridizationeventshavebeenadaptive (vonHoldtetal2016) it is important tonotethatthefullgenome‐wideconsequencesandthegeographicextentofinterspecieshybridizationhavenotbeendocumentedthroughouttheentireeasternrange

Overallourobjectivesweretoquantifygenomicstructureanddiversity across the historical coyote range and the two recentlyexpandedeasterncoyotepopulationsWethenidentifyoutlierlocithatmayhavebeenunderselectioninbothpopulationsasaresultofrangeexpansionWepredictthatpopulationstructurewillcorre‐spond to theknowndemographichistoryofNorthAmericancoy‐otesthatisahistoricalrangepopulationandtwodistinctrecentlyexpandedgroups In accordancewith theoretical assumptionsweexpect reduced genomic diversity in the two recently expandedeastern populations relative to the historical range However hy‐bridizationwith otherCanis speciesmay result in deviations fromthis expectation Finally we expect genomic variants that under‐wentfrequencyshiftsinthesamedirectioninbothgroupstohaveputativefunctionsrelatedtorangeexpansionsuchasdispersalandreproduction

2emsp |emspMETHODS

21emsp|emspSample collection

Weobtainedcoyotebloodand tissue (eg liver kidney tongue)fromstatemanagementprograms(PrincetonIACUC1961A‐13)governmentorganizationarchives(egFloridaFishandWildlifeUS Department of Agriculture Ontario Ministry of NaturalResources and Forestry) or museum archives (New York StateMuseum Oklahoma Museum of Natural History) In all casesstateorprovinceoforiginwasdocumented(Figure1)andinmanycasessexapproximateageandfine‐scalegeographicdatawerealso known Samples were collected between 1998 and 2017withthemajoritycollectedwithinthelast10years(2008ndash2017)Sampleswithunknowncollectiondateswereeitherknownoras‐sumedtofallwithintheapproximatetimeperiod(Supportingin‐formationTableS1)

For downstream analyses we considered samples collectedfrom AZ CA IDMNMO NE NM NV OK SK TXWA andWYtobepartofthehistoricalrange(iepre‐1900Figure1)asdescribed by Hody and Kays (2018) Additionally samples col‐lectedfromMENBNJONandPAwereconsideredpartofthenortheastexpansionandsamplescollectedfromALFLGAKYLANCSCTNandVAwereconsideredpartofthesoutheasternexpansion

22emsp|emspSampling and DNA extraction

HighmolecularweightgenomicDNAwasextractedfromallsampleswitheithertheQiagenDNeasyBloodandTissueKitortheBioSprint96DNABloodKitinconjunctionwithaThermoScientificKingFisherFlexPurificationplatforminbothcasesfollowinginstructionspro‐videdbythemanufacturerWequantifiedDNAconcentrationwitheitherPicoGreenorQubit20 fluorometryand standardized sam‐ples to5ngμlOnlyhigh‐qualityDNAsamples asdeterminedbythepresenceofahighmolecularweightbandona1agarosegelwereretainedforsequencing

23emsp|emspRADsequencing and bioinformatics processing

Wepreparedgenomiclibrariesfollowingamodifiedrestriction‐as‐sociatedDNAsequencing(RADseq)protocoldescribedbyAlietal(2016)BrieflysamplesweredigestedwithsbfIandaunique8bpbarcoded biotinylated adapter was ligated to the resulting frag‐mentsSampleswerethenpooled(96ndash153samplespool)andran‐domlyshearedto400bpinaCovarisLE220FollowingshearingweusedaDynabeadsM‐280streptavidinbeadbindingassaytoenrichforadapter‐ligatedfragmentsFinalsequencinglibrarieswerethenpreparedusingeithertheNEBNextUltraDNALibraryPrepKitortheNEBNextUltraIIDNALibraryPrepKitSizeselectionwasmadefor300ndash400bpinsertwithAgencourtAMPureXPmagneticbeadswhichwerealsousedforlibrarypurificationLibrarieswerestand‐ardized to 10nM and sequenced (2X150nt) on two lanes on theIlluminaHiSeq2500

As thisRADseqprotocol isunique in that thebarcodemaybeoneithertheforwardorreversereaddataprocessingwasrequiredprior to variant calling Accordingly forward and reverse raw se‐quencing readswereprocessedsuch thatany readcontaining theremnantsbfIcutsiteandoneofthepossiblebarcodeswerealignedinasinglefilewhilethematchingreadpairsthatlackedthecutsitewere aligned in a separate file and all remaining reads were dis‐cardedThiswasaccomplishedusingacustomPerlscript(flip_trim_sbfI_170601plseeSupportinginformation)

AdditionaldataprocessingwasthenconductedinSTACKSv142 (CatchenHohenloheBasshamAmoresampCresko2013)First readsweredemultiplexedusingprocess_radtags allowinga 2bpmismatch for barcode rescue and discarding readswitheitheruncalledbasesora low‐qualityscore (lt10)withinaslid‐ing window of 015 Next PCR duplicates were removedwiththe paired‐end sequencing filtering option in clone_filter We

4emsp |emsp emspensp HEPPENHEIMER Et al

excluded all sampleswith low read count (lt500000) after re‐movalofPCRduplicates fromfurtheranalysisRemainingsam‐pleswerethenmappedtothedoggenomeCanFam31assembly(Lindblad‐Tohetal2005)inStampyv1021(LunterampGoodson2011)Toreducesbiasesassociatedwithincorrectlymappedloci(egparalogs)weadditionallyfilteredmappedreadsforamin‐imumMAPQof96andconvertedtobamformat inSamtoolsv0118(Lietal2009)

We completed SNP calling in STACKS following the rec‐ommended pipeline for referencemapped data (iepstacks rarr cstacks rarr sstacks rarr populations Catchenetal2013)Inpstackswe requiredaminimumdepthofcoverageof three to reportastack (‐m3)Cstacks and sstackswere run as recommendedbyCatchenetal (2013)ThepopulationsmodulewasruntwicetooptimizefinalsampleselectiontoreducebothmissingdataandbiasesresultingfromunevensamplingacrosslocationsFirstweallowedonlythefirstSNPperlocus(‐‐write_single_snp)tobere‐portedbutdidnotapplyanymissingdata thresholdsWethenevaluatedtheper‐individualgenotypingsuccessasmeasuredbymissingnessperindividualinPlinkv190b3i(Purcelletal2007)andremovedindividualswithahighlevelofmissingdata(gt80missing)Wealsoremovedsamplesfromlocationswheren = 1 Inoursecondrunofpopulationsweincludedonlythisreducedsetofsamplesrequiredthatreportedlocibegenotypedin90ofindividuals(minusr=09)andagainrestrictedanalysistoonlythe

firstSNPperlocusOnlythislatterdatasetwasusedforsubse‐quentanalysesFollowingSNPcallingwefilteredforstatisticallinkagedisequilibriuminPlinkwiththeargument‐‐indep‐pairwise 50 5 05

All SNPs were annotated as genic (intron or exon) within apromoter(iewithin2Kboftranscriptionstartsitefollowingvon‐HoldtHeppenheimerPetrenkoCroonquistandRutledge(2017))orintergenicusinganin‐housepythonscript(chr_sitepySeesup‐porting information)All intergenic SNPswere compiled in a sec‐ondgenotypedatasetandfilteredforHardyndashWeinbergEquilibrium(HWE) in Plinkwith the argumentmdashhwe 0001 These intergenicHWE‐filtered genotypes were presumed neutral in downstreamanalyses (hereafterputativelyneutral loci)Additionallywecom‐piled a third dataset of putatively functional loci consistingof allSNPsannotatedasgenicorwithin2Kbofatranscriptionstartsite(hereaftergenicloci)

24emsp|emspPopulation structure analysis

Tovisualize clustering inourdata and identify strongoutliersweconductedaPrincipalComponentAnalysisusingourfullSNPdata‐set with flashPCA (Abraham amp Inouye 2014) We identified onestrongoutlieroriginating fromOntariowhichmaybeamisidenti‐fiedeasternorgraywolf (C lycaon or C lupus)This individualwasremovedfromfurtheranalyses

F I G U R E 1 emspMapofcoyotehistoricalrange1950srange2000srangeandsamplesizeperlocationRangesareapproximateandmodifiedfromHodyandKays(2018)

AL GA

FL

SC

LA

1950s2000s

29

13

10

26

12

4

4

11

20

2

5

12

6

4

28

16 23

13

27

25

373

144

Historical range (pre-1900)

14

26

2

4

NCVVAVV

PAPP NJNJNY

MEMENB

ON

TTXX

AZAZ

MNMN

CCAA

WYWWYY

IDD

WWWAAWWW

NVNV

SKSSK

OKOOK

NENEEMOMMOO

KKKKKKKYKTNN

NMNM

emspensp emsp | emsp5HEPPENHEIMER Et al

Following the removal of strong outliers (eg putatively mis‐identified wolves) we determined the most likely number of ge‐nomicclustersrepresentedbythedatabyconductingananalysisofpopulationstructureinADMIXTUREv13(AlexanderNovembreampLange2009)withthecross‐validationflagWeevaluatedK=1ndash10withtheKvaluewiththelowestcross‐validation(cv)scoreindicativeof thebest fitKADMIXTURE is similar inprinciple to theclassicBayesian software STRUCTURE (Pritchard Stephens ampDonnelly2000)butusesamaximumlikelihoodframeworkandismorecom‐putationallyefficientforSNPdata

25emsp|emspGenomic diversity

Standardmetricsofgenomicdiversityforeachsamplinglocationin‐cludingprivateallelecountsaswellasobserved(Ho)andexpected(He) heterozygositywerecalculatedinSTACKSacrossall lociTodeter‐minewhethertrendsofheterozygositydifferedbasedonwhichpartofthegenomewassurveyedwerecalculatedbothHoandHeacrossputativelyneutrallociandgeniclociAllelicrichness(Ar)andprivateallelicrichness(Apr)werecalculatedusingararefactionapproachim‐plementedinADZEv10(SzpiechJakobssonampRosenberg2008)wherethemaximumstandardizedsamplesizewassettothesmallestnforthesamplesconsidered(ie176whencomparingthehistoricalrangenortheastexpansionandsoutheastexpansion)

PairwiseFSTvaluesbetweenall sampling locationsoverall lociwerecalculatedinSTACKSandwetestedforisolationbydistance(IBD)within thecoyotehistorical rangeaswellaswithineachre‐cently expanded eastern populationwith a series ofMantel testsimplementedinade4v17‐11(DrayampDufour2007)inRv33(RCoreTeam2013)PairwiseFSTwerelinearizedfollowingRousset(1997)geographic distanceswere calculated as the shortest straight‐linedistancebetweensamplinglocationsandsignificancewasassessedfrom9999permutations

26emsp|emspIdentification of loci associated with range expansion

Toidentifylociascandidatesforselectionduringrangeexpansionwerestrictedanalysestoindividualswithhighclusterassignmentsas identified intheADMIXTUREanalysis (Qge08)Additionallytopreventbiasesresultingfromlong‐distancedispersersweremovedthreeindividualsthathadhighclusterassignmentstodifferentpop‐ulationsfromwhichtheyweresampled(egsampledfromLouisianabut clustered with the northeast group) Though restricting theanalysestoindividualswithhighclusterassignmentsmayinflatethegeneticdistinctionbetweenthehistoricalrangeandtherecentlyex‐pandedcoyotepopulationswebelievetheinferredhistoricalrangepopulationtobethebestavailablerepresentationofpre‐rangeex‐pansionallelefrequenciesasthereis likelyongoingcontemporarygeneflowbetweencoyotesinthehistoricalrangeandbothrecentlyexpandedpopulations

AsFSToutlier‐typeapproachestoidentifylociunderselectionarepronetohighratesoffalsepositivesespeciallyamongpopulations

with complex demographic histories we used two distinct ap‐proachestoidentifylociputativelyunderselectionFirstaBayesianframework to detect outlier loci was implemented in BAYENV2(CoopWitonskyRienzoampPritchard2010GuumlntherampCoop2013)Thismethodaccountsforevolutionarynonindependencebetweenpopulationsbyfirstcalculatingcovarianceinallelefrequenciesataset of putatively neutral loci Candidate functional SNPs are thenevaluatedoneata timeunderamodel thatassumesa linear rela‐tionship between an environmental variable and allele frequencycomparedtoamodelgivenbytheneutralcovariancematrixandacorrespondingBayesfactoriscalculatedThismethodhasbeensug‐gested to outperformotherFSToutlier‐likemethods (eg FDIST2BayeScan) in the case of range expansion (Lotterhos ampWhitlock2014) In theBAYENV2analysis theenvironmentalvariableof in‐terestwasthelineardistancefromthecoyotehistoricalrange(egWhiteetal2013)Toavoidbiasesinducedbyallelefrequenciesatanyonesampling locationwithin thehistorical rangeall statesorprovinceswithinthehistoricalcoyoterangeweretreatedasasinglesampling locationDistances for sampling locationsoutsideof thehistoricalrangewerecalculatedastheshorteststraight‐linedistancefrom theapproximatemidpointof the sampled regionswithin thehistoricalrangeThenorthernandsouthernexpansionfrontswereanalyzedseparately Inbothcasestheputativelyneutral lociusedtogeneratethecontrolcovariancematrixwere intergenicSNPs inHWEfilteredfurthertoremoveanymonomorphiclocibetweenthepopulationscomparedSimilarlythecandidateSNPswereallSNPsannotatedasgenic(intronorexon)orwithin2Kbofatranscriptionstartsiteagainfilteredtoremovemonomorphiclocibetweenpop‐ulationsGenotype files forbothSNPdatasetswereconverted toBAYENV2formatinPGDSpiderv2113(LischerampExcoffier2012)For each expansion front (historical to northeast amp historical tosoutheast)lociwererankedbyBayesfactorandthetop3ofSNPswereretainedforfurtheranalysisLociwereconsideredcandidategenesunderselectionduringrangeexpansionifthesameSNPoc‐curredonbothlists

Secondweusedaprincipalcomponent‐basedapproachimple‐mentedwith theRpackagePCadapt (LuuBazinampBlum2017)This method first performs a centered scaled principal compo‐nentanalysisongenome‐wideSNPsandthenidentifiessignificantoutlierswith respect topopulation structuregivenby the firstK principal components Specifically PCadapt identifies outliersbasedon theMahalanobisdistancewhichdescribesmultidimen‐sionaldistanceofapointfromthemeanSimulationsindicatethatPCadaptislesspronetotypeIIerrorthanalternativemethods(egBayeScan)andPCadaptisexpectedtoperformwellunderavarietyofcomplexdemographicscenariosincludingrangeexpansion(Luuetal2017)

InthePCadaptanalysisthenortheasternandsoutheasternex‐pansionfrontswereanalyzedseparatelyInbothcasesweidenti‐fiedoutlierswithrespecttounderlyingpopulationstructuregivenby PC1We chose to retain only PC1 rather than selecting theoptimalnumberofPCsbasedonconventionalmethods(egscreeplot)asPC1primarilycapturedthemajoraxisofrangeexpansion

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 4: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

4emsp |emsp emspensp HEPPENHEIMER Et al

excluded all sampleswith low read count (lt500000) after re‐movalofPCRduplicates fromfurtheranalysisRemainingsam‐pleswerethenmappedtothedoggenomeCanFam31assembly(Lindblad‐Tohetal2005)inStampyv1021(LunterampGoodson2011)Toreducesbiasesassociatedwithincorrectlymappedloci(egparalogs)weadditionallyfilteredmappedreadsforamin‐imumMAPQof96andconvertedtobamformat inSamtoolsv0118(Lietal2009)

We completed SNP calling in STACKS following the rec‐ommended pipeline for referencemapped data (iepstacks rarr cstacks rarr sstacks rarr populations Catchenetal2013)Inpstackswe requiredaminimumdepthofcoverageof three to reportastack (‐m3)Cstacks and sstackswere run as recommendedbyCatchenetal (2013)ThepopulationsmodulewasruntwicetooptimizefinalsampleselectiontoreducebothmissingdataandbiasesresultingfromunevensamplingacrosslocationsFirstweallowedonlythefirstSNPperlocus(‐‐write_single_snp)tobere‐portedbutdidnotapplyanymissingdata thresholdsWethenevaluatedtheper‐individualgenotypingsuccessasmeasuredbymissingnessperindividualinPlinkv190b3i(Purcelletal2007)andremovedindividualswithahighlevelofmissingdata(gt80missing)Wealsoremovedsamplesfromlocationswheren = 1 Inoursecondrunofpopulationsweincludedonlythisreducedsetofsamplesrequiredthatreportedlocibegenotypedin90ofindividuals(minusr=09)andagainrestrictedanalysistoonlythe

firstSNPperlocusOnlythislatterdatasetwasusedforsubse‐quentanalysesFollowingSNPcallingwefilteredforstatisticallinkagedisequilibriuminPlinkwiththeargument‐‐indep‐pairwise 50 5 05

All SNPs were annotated as genic (intron or exon) within apromoter(iewithin2Kboftranscriptionstartsitefollowingvon‐HoldtHeppenheimerPetrenkoCroonquistandRutledge(2017))orintergenicusinganin‐housepythonscript(chr_sitepySeesup‐porting information)All intergenic SNPswere compiled in a sec‐ondgenotypedatasetandfilteredforHardyndashWeinbergEquilibrium(HWE) in Plinkwith the argumentmdashhwe 0001 These intergenicHWE‐filtered genotypes were presumed neutral in downstreamanalyses (hereafterputativelyneutral loci)Additionallywecom‐piled a third dataset of putatively functional loci consistingof allSNPsannotatedasgenicorwithin2Kbofatranscriptionstartsite(hereaftergenicloci)

24emsp|emspPopulation structure analysis

Tovisualize clustering inourdata and identify strongoutliersweconductedaPrincipalComponentAnalysisusingourfullSNPdata‐set with flashPCA (Abraham amp Inouye 2014) We identified onestrongoutlieroriginating fromOntariowhichmaybeamisidenti‐fiedeasternorgraywolf (C lycaon or C lupus)This individualwasremovedfromfurtheranalyses

F I G U R E 1 emspMapofcoyotehistoricalrange1950srange2000srangeandsamplesizeperlocationRangesareapproximateandmodifiedfromHodyandKays(2018)

AL GA

FL

SC

LA

1950s2000s

29

13

10

26

12

4

4

11

20

2

5

12

6

4

28

16 23

13

27

25

373

144

Historical range (pre-1900)

14

26

2

4

NCVVAVV

PAPP NJNJNY

MEMENB

ON

TTXX

AZAZ

MNMN

CCAA

WYWWYY

IDD

WWWAAWWW

NVNV

SKSSK

OKOOK

NENEEMOMMOO

KKKKKKKYKTNN

NMNM

emspensp emsp | emsp5HEPPENHEIMER Et al

Following the removal of strong outliers (eg putatively mis‐identified wolves) we determined the most likely number of ge‐nomicclustersrepresentedbythedatabyconductingananalysisofpopulationstructureinADMIXTUREv13(AlexanderNovembreampLange2009)withthecross‐validationflagWeevaluatedK=1ndash10withtheKvaluewiththelowestcross‐validation(cv)scoreindicativeof thebest fitKADMIXTURE is similar inprinciple to theclassicBayesian software STRUCTURE (Pritchard Stephens ampDonnelly2000)butusesamaximumlikelihoodframeworkandismorecom‐putationallyefficientforSNPdata

25emsp|emspGenomic diversity

Standardmetricsofgenomicdiversityforeachsamplinglocationin‐cludingprivateallelecountsaswellasobserved(Ho)andexpected(He) heterozygositywerecalculatedinSTACKSacrossall lociTodeter‐minewhethertrendsofheterozygositydifferedbasedonwhichpartofthegenomewassurveyedwerecalculatedbothHoandHeacrossputativelyneutrallociandgeniclociAllelicrichness(Ar)andprivateallelicrichness(Apr)werecalculatedusingararefactionapproachim‐plementedinADZEv10(SzpiechJakobssonampRosenberg2008)wherethemaximumstandardizedsamplesizewassettothesmallestnforthesamplesconsidered(ie176whencomparingthehistoricalrangenortheastexpansionandsoutheastexpansion)

PairwiseFSTvaluesbetweenall sampling locationsoverall lociwerecalculatedinSTACKSandwetestedforisolationbydistance(IBD)within thecoyotehistorical rangeaswellaswithineachre‐cently expanded eastern populationwith a series ofMantel testsimplementedinade4v17‐11(DrayampDufour2007)inRv33(RCoreTeam2013)PairwiseFSTwerelinearizedfollowingRousset(1997)geographic distanceswere calculated as the shortest straight‐linedistancebetweensamplinglocationsandsignificancewasassessedfrom9999permutations

26emsp|emspIdentification of loci associated with range expansion

Toidentifylociascandidatesforselectionduringrangeexpansionwerestrictedanalysestoindividualswithhighclusterassignmentsas identified intheADMIXTUREanalysis (Qge08)Additionallytopreventbiasesresultingfromlong‐distancedispersersweremovedthreeindividualsthathadhighclusterassignmentstodifferentpop‐ulationsfromwhichtheyweresampled(egsampledfromLouisianabut clustered with the northeast group) Though restricting theanalysestoindividualswithhighclusterassignmentsmayinflatethegeneticdistinctionbetweenthehistoricalrangeandtherecentlyex‐pandedcoyotepopulationswebelievetheinferredhistoricalrangepopulationtobethebestavailablerepresentationofpre‐rangeex‐pansionallelefrequenciesasthereis likelyongoingcontemporarygeneflowbetweencoyotesinthehistoricalrangeandbothrecentlyexpandedpopulations

AsFSToutlier‐typeapproachestoidentifylociunderselectionarepronetohighratesoffalsepositivesespeciallyamongpopulations

with complex demographic histories we used two distinct ap‐proachestoidentifylociputativelyunderselectionFirstaBayesianframework to detect outlier loci was implemented in BAYENV2(CoopWitonskyRienzoampPritchard2010GuumlntherampCoop2013)Thismethodaccountsforevolutionarynonindependencebetweenpopulationsbyfirstcalculatingcovarianceinallelefrequenciesataset of putatively neutral loci Candidate functional SNPs are thenevaluatedoneata timeunderamodel thatassumesa linear rela‐tionship between an environmental variable and allele frequencycomparedtoamodelgivenbytheneutralcovariancematrixandacorrespondingBayesfactoriscalculatedThismethodhasbeensug‐gested to outperformotherFSToutlier‐likemethods (eg FDIST2BayeScan) in the case of range expansion (Lotterhos ampWhitlock2014) In theBAYENV2analysis theenvironmentalvariableof in‐terestwasthelineardistancefromthecoyotehistoricalrange(egWhiteetal2013)Toavoidbiasesinducedbyallelefrequenciesatanyonesampling locationwithin thehistorical rangeall statesorprovinceswithinthehistoricalcoyoterangeweretreatedasasinglesampling locationDistances for sampling locationsoutsideof thehistoricalrangewerecalculatedastheshorteststraight‐linedistancefrom theapproximatemidpointof the sampled regionswithin thehistoricalrangeThenorthernandsouthernexpansionfrontswereanalyzedseparately Inbothcasestheputativelyneutral lociusedtogeneratethecontrolcovariancematrixwere intergenicSNPs inHWEfilteredfurthertoremoveanymonomorphiclocibetweenthepopulationscomparedSimilarlythecandidateSNPswereallSNPsannotatedasgenic(intronorexon)orwithin2Kbofatranscriptionstartsiteagainfilteredtoremovemonomorphiclocibetweenpop‐ulationsGenotype files forbothSNPdatasetswereconverted toBAYENV2formatinPGDSpiderv2113(LischerampExcoffier2012)For each expansion front (historical to northeast amp historical tosoutheast)lociwererankedbyBayesfactorandthetop3ofSNPswereretainedforfurtheranalysisLociwereconsideredcandidategenesunderselectionduringrangeexpansionifthesameSNPoc‐curredonbothlists

Secondweusedaprincipalcomponent‐basedapproachimple‐mentedwith theRpackagePCadapt (LuuBazinampBlum2017)This method first performs a centered scaled principal compo‐nentanalysisongenome‐wideSNPsandthenidentifiessignificantoutlierswith respect topopulation structuregivenby the firstK principal components Specifically PCadapt identifies outliersbasedon theMahalanobisdistancewhichdescribesmultidimen‐sionaldistanceofapointfromthemeanSimulationsindicatethatPCadaptislesspronetotypeIIerrorthanalternativemethods(egBayeScan)andPCadaptisexpectedtoperformwellunderavarietyofcomplexdemographicscenariosincludingrangeexpansion(Luuetal2017)

InthePCadaptanalysisthenortheasternandsoutheasternex‐pansionfrontswereanalyzedseparatelyInbothcasesweidenti‐fiedoutlierswithrespecttounderlyingpopulationstructuregivenby PC1We chose to retain only PC1 rather than selecting theoptimalnumberofPCsbasedonconventionalmethods(egscreeplot)asPC1primarilycapturedthemajoraxisofrangeexpansion

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 5: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp5HEPPENHEIMER Et al

Following the removal of strong outliers (eg putatively mis‐identified wolves) we determined the most likely number of ge‐nomicclustersrepresentedbythedatabyconductingananalysisofpopulationstructureinADMIXTUREv13(AlexanderNovembreampLange2009)withthecross‐validationflagWeevaluatedK=1ndash10withtheKvaluewiththelowestcross‐validation(cv)scoreindicativeof thebest fitKADMIXTURE is similar inprinciple to theclassicBayesian software STRUCTURE (Pritchard Stephens ampDonnelly2000)butusesamaximumlikelihoodframeworkandismorecom‐putationallyefficientforSNPdata

25emsp|emspGenomic diversity

Standardmetricsofgenomicdiversityforeachsamplinglocationin‐cludingprivateallelecountsaswellasobserved(Ho)andexpected(He) heterozygositywerecalculatedinSTACKSacrossall lociTodeter‐minewhethertrendsofheterozygositydifferedbasedonwhichpartofthegenomewassurveyedwerecalculatedbothHoandHeacrossputativelyneutrallociandgeniclociAllelicrichness(Ar)andprivateallelicrichness(Apr)werecalculatedusingararefactionapproachim‐plementedinADZEv10(SzpiechJakobssonampRosenberg2008)wherethemaximumstandardizedsamplesizewassettothesmallestnforthesamplesconsidered(ie176whencomparingthehistoricalrangenortheastexpansionandsoutheastexpansion)

PairwiseFSTvaluesbetweenall sampling locationsoverall lociwerecalculatedinSTACKSandwetestedforisolationbydistance(IBD)within thecoyotehistorical rangeaswellaswithineachre‐cently expanded eastern populationwith a series ofMantel testsimplementedinade4v17‐11(DrayampDufour2007)inRv33(RCoreTeam2013)PairwiseFSTwerelinearizedfollowingRousset(1997)geographic distanceswere calculated as the shortest straight‐linedistancebetweensamplinglocationsandsignificancewasassessedfrom9999permutations

26emsp|emspIdentification of loci associated with range expansion

Toidentifylociascandidatesforselectionduringrangeexpansionwerestrictedanalysestoindividualswithhighclusterassignmentsas identified intheADMIXTUREanalysis (Qge08)Additionallytopreventbiasesresultingfromlong‐distancedispersersweremovedthreeindividualsthathadhighclusterassignmentstodifferentpop‐ulationsfromwhichtheyweresampled(egsampledfromLouisianabut clustered with the northeast group) Though restricting theanalysestoindividualswithhighclusterassignmentsmayinflatethegeneticdistinctionbetweenthehistoricalrangeandtherecentlyex‐pandedcoyotepopulationswebelievetheinferredhistoricalrangepopulationtobethebestavailablerepresentationofpre‐rangeex‐pansionallelefrequenciesasthereis likelyongoingcontemporarygeneflowbetweencoyotesinthehistoricalrangeandbothrecentlyexpandedpopulations

AsFSToutlier‐typeapproachestoidentifylociunderselectionarepronetohighratesoffalsepositivesespeciallyamongpopulations

with complex demographic histories we used two distinct ap‐proachestoidentifylociputativelyunderselectionFirstaBayesianframework to detect outlier loci was implemented in BAYENV2(CoopWitonskyRienzoampPritchard2010GuumlntherampCoop2013)Thismethodaccountsforevolutionarynonindependencebetweenpopulationsbyfirstcalculatingcovarianceinallelefrequenciesataset of putatively neutral loci Candidate functional SNPs are thenevaluatedoneata timeunderamodel thatassumesa linear rela‐tionship between an environmental variable and allele frequencycomparedtoamodelgivenbytheneutralcovariancematrixandacorrespondingBayesfactoriscalculatedThismethodhasbeensug‐gested to outperformotherFSToutlier‐likemethods (eg FDIST2BayeScan) in the case of range expansion (Lotterhos ampWhitlock2014) In theBAYENV2analysis theenvironmentalvariableof in‐terestwasthelineardistancefromthecoyotehistoricalrange(egWhiteetal2013)Toavoidbiasesinducedbyallelefrequenciesatanyonesampling locationwithin thehistorical rangeall statesorprovinceswithinthehistoricalcoyoterangeweretreatedasasinglesampling locationDistances for sampling locationsoutsideof thehistoricalrangewerecalculatedastheshorteststraight‐linedistancefrom theapproximatemidpointof the sampled regionswithin thehistoricalrangeThenorthernandsouthernexpansionfrontswereanalyzedseparately Inbothcasestheputativelyneutral lociusedtogeneratethecontrolcovariancematrixwere intergenicSNPs inHWEfilteredfurthertoremoveanymonomorphiclocibetweenthepopulationscomparedSimilarlythecandidateSNPswereallSNPsannotatedasgenic(intronorexon)orwithin2Kbofatranscriptionstartsiteagainfilteredtoremovemonomorphiclocibetweenpop‐ulationsGenotype files forbothSNPdatasetswereconverted toBAYENV2formatinPGDSpiderv2113(LischerampExcoffier2012)For each expansion front (historical to northeast amp historical tosoutheast)lociwererankedbyBayesfactorandthetop3ofSNPswereretainedforfurtheranalysisLociwereconsideredcandidategenesunderselectionduringrangeexpansionifthesameSNPoc‐curredonbothlists

Secondweusedaprincipalcomponent‐basedapproachimple‐mentedwith theRpackagePCadapt (LuuBazinampBlum2017)This method first performs a centered scaled principal compo‐nentanalysisongenome‐wideSNPsandthenidentifiessignificantoutlierswith respect topopulation structuregivenby the firstK principal components Specifically PCadapt identifies outliersbasedon theMahalanobisdistancewhichdescribesmultidimen‐sionaldistanceofapointfromthemeanSimulationsindicatethatPCadaptislesspronetotypeIIerrorthanalternativemethods(egBayeScan)andPCadaptisexpectedtoperformwellunderavarietyofcomplexdemographicscenariosincludingrangeexpansion(Luuetal2017)

InthePCadaptanalysisthenortheasternandsoutheasternex‐pansionfrontswereanalyzedseparatelyInbothcasesweidenti‐fiedoutlierswithrespecttounderlyingpopulationstructuregivenby PC1We chose to retain only PC1 rather than selecting theoptimalnumberofPCsbasedonconventionalmethods(egscreeplot)asPC1primarilycapturedthemajoraxisofrangeexpansion

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 6: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

6emsp |emsp emspensp HEPPENHEIMER Et al

and therefore corresponded to the level of population structurerelevant for this study (See Figure 2) To evaluate significancep‐values foreachSNPweretransformed intoq‐valuesandSNPswithq‐valueslt005wereretainedthereforecontrollingforafalsediscoveryrate(FDR)of5ThiswasimplementedintheRpack‐ageqvaluev26(BassDabneyampRobinson2018)AgainweonlyconsideredSNPsthatweresignificantinbothhistoricalandnorth‐east and historical and southeast comparisons as candidates forlociunderselectionduringrangeexpansion

Genefunctionsandgeneontologybiologicalprocessannota‐tionsofalloutlierSNPsidentifiedbyboththeanalysesmethodswereinferredusingEnsembl(release91Zerbinoetal2017)andAmiGO 2 accessed in February 2018 (Carbon et al 2009)Weconductedageneontology(GO)biologicalprocessoverrepresen‐tation enrichment analysis on outlier sites located in functionalgenomicregions(ieintronexonorpromoter)usingWebGestalt(ZhangKirovampSnoddy2005Wangetal2013)WeusedourgenicSNPdatasetas the referenceset for theenrichmentanal‐ysis and significance was evaluated using an FDR threshold of5 Additionally we used the Ensembl Variant Effect Predictor(McLarenetal2016)topredictthefunctionaleffectofalloutlierSNPs

3emsp |emspRESULTS

31emsp|emspGenotyping

Among the final samples retained for analyses raw sequencingreadcountspersamplerangedbetween770960and13299663withanaverageof2466167readsFiltering forPCRduplicatesprior to SNP calling removedbetween532and5615 (aver‐age 2148 Supporting information Table S2) of reads leav‐ing between 582946 and 10786513 (average 1908634Supporting informationTable S2) reads assigned to eachuniquebarcodeMappability to the dog genome following the removalof readswith lowMAPQscores ranged from6377 to7920(average7404)

Wesequencedatotalof3597305restriction‐associatedsites(RADtags)24139ofwhichcontainedatleastonevariablesitein394coyotesFollowingLDfilteringourfulldatasetconsistedof22935 biallelic SNP loci with a total genotyping rate of 933(Supporting information Figure S1) Average per‐individualmiss‐ingnesswashighestinthenortheast(meanmissing=105)andslightlylowerinboththesoutheast(60)andthehistoricalrange(54SupportinginformationFigureS2)Averagedepthofcover‐ageacrossallindividualswas11333withastandarddeviationof6626andwassimilaramongthethreeregionssurveyed(histori‐calrange11420stdev=5647northeast11634stdev=9991southeast 11096 stdev=4988 Supporting informationFigureS3)Furthermoreoverallallelebalanceforallheterozygotes(ieminor allele coverage relative to total site coverage) was 0496(stdev=0113)andagainsimilaracrossall three regions (histori‐cal range 0496 stdev=0113 northeast 0493 stdev=0114southeast0497stdev=0112)

Overall we primarily captured rare variation with an averageglobalminorallelefrequencyof173Furtherwithineachofthethreeregionssampledminorallelefrequenciesweretypicallyle5(SupportinginformationFigureS4)Approximatelyhalfofthesiteswereintergenic(nintergenic=12676)with14108SNPsfoundwithingenes(nintron=12024nexon=1532npromoter=552)Theseannota‐tionssumtogt22935asSNPsmayhavemultipleannotations(egpromoter and intron) Additionally our putatively neutral datasetwhichconsistedofintergenicSNPsinHWEretained11518SNPsOur genic data included 10259 SNPs within introns exons andpromoters

32emsp|emspPopulation structure corresponds to expansion axis

Our PCA divided sampling locations as predicted with PC1(111 variance explained) separating samples originating fromthe historical coyote range fromeither recently expanded east‐ern population (Figure 2a) Accordingly PC1 was significantlycorrelatedwiththelongitudeofsamplinglocation(stateorprov‐ince Pearsonrsquos r = 091 plt22times10minus16 Figure 2b) PC2 (086variance explained) primarily separated northeastern sampling

F I G U R E 2 emsp (a)Principalcomponentanalysis(PCA)ofmoleculardataforall394coyotesInsertGeographicdistributionofsamplinglocations(b)CorrelationbetweenPC1andlongitudeofsamplinglocation(Pearsonsr = 091 plt22times10minus16)

minus40 minus20 0 20 40 60 80

PC1 (111)

Long

itude

r = 091

minus40 minus20 0 20 40 60 80

minus20

0

20

40

PC1 (111)

PC2

(08

6)

Historical rangeSoutheast expansionNortheast expansion

Historical rangeSoutheast expansionNortheast expansion

(a)

(b)

7080

9010

011

012

0

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 7: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp7HEPPENHEIMER Et al

locations from southeastern sampling locations Furthermoresamples fromtheMid‐Atlanticcontactzonebetweenthesetwofrontsofexpansion(NorthCarolinaVirginia)tendedtohavein‐termediatespatialplacementonPC2(Figure2a)

Our ADMIXTURE analysis indicated that three distinct ge‐netic clusters were best represented by the data (cv=0107Figure 3 Supporting information Figure S5) Generally theseclusters were concordant with sampling location with onecluster corresponding to the historical coyote range a secondclustercorresponding to thenortheasternexpansion frontanda third cluster corresponding to the southeastern expansion(Figure3b)thatissamplescollectedfromthehistoricalcoyoterange showed high assignments to the historical range cluster(Average QHistorical=0868 Supporting information Table S3)andsimilarlysamplesobtainedfromeitherthenortheasternorsoutheastern expansion front were strongly assigned to eachrespectiveclusteraverageQNortheast=0943andsoutheastav‐erageQSoutheast=0933 (Supporting information Table S3) Weobservedmoderatelyhighfrequenciesof intermediateancestryassignmentstobothrecentlyexpandedeasternpopulationintheMid‐Atlanticregion(egNCKYVAPAFigure3bSupportinginformation Table S3) consistent with this region as the loca‐tion of recent secondary contact between the two expansionfronts(Heppenheimeretal2018)Despitethecleanseparationofclustersbasedonknownexpansion routesonesampleorig‐inatingfromLouisianastronglyclusteredwiththenortheasterngroupAdditionallysomesamplinglocationswithinthehistoricalrange(egMOOKNEMN)exhibitedintermediateassignmentstothesoutheasternclusterFurthermoreclusteringatK=2wasconsistentwithonehistoricalrangeclusterandonerecentlyex‐pandedgroup(Figure3a)

33emsp|emspHigh genomic diversity in recently expanded populations

Genome‐wide heterozygosity was approximately equivalent acrossthehistorical range (AverageHE=00264)andnortheasternexpan‐sion front (AverageHE=00268)andslightlyelevated in thesouth‐easternexpansionfront(AverageHE=00300)Theserelativetrendsweresimilarwhenanalysiswasrestrictedtoeitherputativelyneutralloci(AverageHEHistorical=00208AverageHENortheast=00195Average HE Southeast=00235 Supporting information TableS4) or genic SNPs (Average HE Historical=00256 Average HE Northeast=00267AverageHESoutheast=00297Supporting in‐formationTableS4)Furthermoreallelicrichnesswashighestinthehistorical rangeand (historicalAr=1489 stderr=0003Figure4a)lowerinbothexpansionfronts(southeastAr=1467stderr=0003northeastAr=1425stderr=0003Figure4a)Privateallelecounts(Table 1) and private allelic richness (Figure 4b) exhibited a simi‐lar trendwith thehighest valuesobserved for thehistorical range(historical Apr=0189 stderr=0002 count=5799) and lowervalues observed in the southeastern front (southeast Apr=0117stderr=0002 count=3578) and northeastern expansion front(northeastApr=0138stderr=0002count=3018)

Wefoundnoevidenceof isolationbydistanceinthehistoricalcoyoterange(MantelR=0154p=0152)orineitherrecentlyex‐panded eastern population (northeast Mantel R=minus0288 0715southeastMantelR=0846p=0327)

34emsp|emspOutlier loci associated with range expansion

With the BAYENV2 approach the Bayes factors for the top 3of ranked SNPs ranged 2080ndash53564 (mean=355825) for the

F I G U R E 3 emspPercentancestryassignments(Q)atK=2(a)andK=3(b)intheadmixtureanalysis

30

40

50

60

70

80

WA CA NV ID AZ WY SK NM NE TX OK MN LA FL AL GA SC NC TN KY VA PA

0

10

20

30

40

50

60

70

80

90

100

0

10

20

90

100

NJ NY ME NB ONMO

Historical range Southeast Northeast

K = 2

K = 3

(a)

(b)

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 8: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

8emsp |emsp emspensp HEPPENHEIMER Et al

northeast and historical range analysis and 1310ndash130670000(mean271374343) for the southeast and historical range analy‐sis (Table2 Supporting informationTableS5)A totalof53genicSNPswere sharedamong the top3of rankedSNPs inboth thenortheastexpansionandhistorical rangeandsoutheastexpansionandhistoricalrangeanalyses(Figure5a)TheseSNPswereprimarilyintronic (nintron=45)withsixexonicSNPsandtwo located inpu‐tativepromoterregions(Table2SupportinginformationTableS5)WiththePCadaptapproach59SNPsweresignificantoutliers(FDR5)inboththenortheastexpansionandhistoricalrangeandsouth‐eastexpansionandhistorical rangeanalyses (Figure5a)Of these22SNPswerewithingenes (nintron = 20 npromoter=2)and37wereintergenic(Table2SupportinginformationTableS5)ThoughthesetwoanalysesmethodstoidentifyoutlierSNPsaresimilarBAYENV2is based on a priori defined groups (ie sampling location) whilePCadaptisbasedonprincipalcomponentscoreswithoutpredefinedgroupsHoweverasPC1washighlycorrelatedwiththeexpansionaxis (seeFigure2b)therewasnodiscordancebetweenplacementalongPC1andthecategorizationofsamplesaseitherhistoricalorrecentlyexpandedAccordinglyweconsiderthisanalysistobedi‐rectlycomparableandfocusourresultsanddiscussiononSNPsandgenesidentifiedbybothanalyses(LotterhosampWhitlock2015)

AtotaloftwelveSNPsof22935wereoutliersinboththeout‐lieranalyses(Table2Figure5a)Inallcasesthechangeinallele

frequencyrelative to thehistorical rangewas in thesamedirec‐tion forboth recentlyexpandedeasternpopulations (Figures5band6)Fornineoftheseoutlierlocithelesscommonalleleoverall(ie theglobalminorallele)decreased intherecentlyexpandedpopulationsandfortheremainingthreelocitheminorallelein‐creasedinfrequency(Figures5band6)InonecaseWDR17theminorallelewaslostinbothoftherecentlyexpandedeasternpop‐ulations(Figure6)Forthreeadditionaloutlierloci(PAX5EPHA6 and CARMIL1) the minor allele was lost in one of the recentlyexpanded populations and substantially reduced in frequencyin the other (Figure 5b) Furthermore therewas only one locus(ZDHHC16) where theminor allelewas absent from the histori‐calrangebutpresentatappreciablefrequenciesinbothrecentlyexpanded populations (qNortheast=1212 qSoutheast=2409Figure6)

InourGOenrichmentanalysistheseoutlierSNPswerenotsig‐nificantlyenrichedforanybiologicalprocess (Supporting informa‐tionTableS6)afterapplyinganFDRthresholdof5FurthermoreallsiteswereannotatedasldquomodifiersrdquointheVEPanalysiswhicharedefinedasvariantswithnopredictablefunctionaleffectsoncodingregions(ienoncodingvariants)

4emsp |emspDISCUSSION

RecentlyexpandedcoyotepopulationsineasternNorthAmericaprovide a unique opportunity to explore how range expansionshapesgenomicdiversityatneutralandputativelyadaptive lociHereweidentifiedthreegeneticgroupsofcoyoteswhichlargelycorrespond to the historical range and two distinct expansionfronts Instances of discordance between sampling location andclusterassignmentswererelativelyrareandlikelyduetorecentsharedancestryaswell asongoinggene flow Inparticular coy‐otesfromOKNEandMNexhibitedintermediateassignmentstothe southeastern cluster and coyotes fromMO exhibited inter‐mediateassignmentstoboththesoutheasternandnortheasternclustersWith the exceptionofMN thismidwestern regionhaspreviouslybeensuggestedtorepresentthesourcepopulationforthesoutheasternexpansion(Nowak1979vonHoldtetal2011)Additionallyascoyotesarehighlymobilethere is likelyongoinggene flow among the recently expanded populations and thosein the historical range particularly along the eastern extremeTo directly address the relative impacts of shared ancestry and ongoinggeneflowintheeasternhistoricalrangeamoreextensivesamplingschemethroughoutthisregionisneededandpre‐rangeexpansionsamplesfrompriorto1900shouldalsobeincluded

AsinHeppenheimeretal(2018)weobservedcomparativelyhigh levels of diversity in both the northeastern and southeast‐ern coyote populationswhich is inconsistentwith the theoreti‐cal(Excoffieretal2009)andempiricalexpectationsforrecentlyexpandedpopulationsForexample recentstudies reportedde‐creased heterozygosity for populations of bank voles (Myodoes glareolus)anddamselflies(Coenagrion scitulum)inexpansionfronts

F I G U R E 4 emspAllelicrichness(a)andprivateallelicrichness(b)acrossalllociasafunctionofstandardizedsamplesize

0 100 150

Sample size (g)

(a)

50

11

12

13

14

15A

llelic

rich

ness

Historical rangeSoutheast expansionNortheast expansion

(b)

0 20 40 60 80 100 120

005

010

015

020

Sample size (g)

Priv

ate

alle

lic ri

chne

ss

Historical rangeSoutheast expansionNortheast expansion

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 9: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp9HEPPENHEIMER Et al

relativetotheirsourcepopulations(Whiteetal2013Swaegerset al 2015) In contrastweobserved approximately equivalentheterozygositybetweencoyotepopulations in thehistoricalandrecently established northeastern ranges andmore intriguinglywe observed that coyotes in the southeastern US had slightlygreaterheterozygositythanthose inthehistoricalrangeAsthistrendisconsistentacrossvarioussubsectionsofthegenome(iegenic and putatively neutral regions) it is not immediately clearfromthisstudywhat isdriving this lackof reduction ingenomic

diversityintherecentlyestablishedeasternpopulationsHoweverthe observed trends for private allele counts and private allelicrichness incoyotesareconsistentwithabottleneckscenarioasthehistoricalrangewasobservedtohavethehighestnumberofprivateallelesandthehighestprivateallelicrichnessAsthelossof rare alleleswill haveagreater impactonallelic richness thanheterozygosity(Greenbaumetal2014)allelicrichnesshasbeensuggestedtobeabetterindicatorofabottleneckparticularlyfol‐lowingrangeexpansion

Sampling locations n Private alleles Ho He

HistoricalRange

Arizona(AZ) 12 470 00254 00278

California(CA) 29 895 00225 00276

Idaho(ID) 10 167 00208 00237

Minnesota(MN) 12 311 00326 00324

Missouri(MO) 6 105 00199 00227

Nebraska(NE) 20 583 00250 00278

NewMexico(NM) 11 401 00276 00288

Nevada(NV) 13 406 00244 00272

Oklahoma(OK) 5 162 00296 00289

Saskatchewan(SK) 4 132 0024 00244

Texas(TX) 2 146 00281 00225

Washington(WA) 14 227 00260 00278

Wyoming(WY) 4 115 00214 00220

Overallhistoricalrange 142 5799a 00252 00264

Northeastexpansion

Maine(ME) 14 499 00231 00292

NewBrunswick(NB) 4 107 00207 00228

NewJersey(NJ) 3 47 00259 00233

NewYork(NY) 4 65 00257 00222

Ontario(ON) 26 464 00305 00324

Pennsylvania(PA) 37 1511 00233 00307

OverallNortheastExpansion 88 3018a 00249 00268

Southeastexpansion

Alabama(AL) 16 136 00319 00326

Florida(FL) 28 843 00262 00311

Georgia(GA) 23 151 00300 00318

Kentucky(KY) 26 280 00298 00318

Louisiana(LA) 4 125 00268 00284

NorthCarolina(NC) 27 406 00296 00321

SouthCarolina(SC) 13 100 00325 00324

Tennessee(TN) 2 32 00295 00228

Virginia(VA) 25 305 00221 00270

OverallSoutheastExpansion 164 3578a 00287 00300

Note nsamplesizeHoobservedheterozygosityHeexpectedheterozygosityaPrivateallelecountsperstateprovincearenotexpectedtosumtotheoverallprivateallelecountper region as allelesmaybeprivate to a regionwithoutbeingprivate to any individual stateorprovince

TA B L E 1 emspSummarystatisticsforallsamplinglocations(n=394)across22935biallelicSNPs

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 10: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

10emsp |emsp emspensp HEPPENHEIMER Et al

Themaintenanceof relativelyhighheterozygosity in recentlyexpandedpopulations could be the result of a selective processsuch as balancing selection along the expansion axes Howeverit is perhapsmore likely that this pattern results from extensivegene floweitherdueto long‐distancedispersalbycoyotesorigi‐natingfromthehistoricalrangeorduetointerbreedingwithotherCanisspeciesinhabitingtheeasternUnitedStatesorsoutheasternCanadaWhilecharacterizinginterspecifichybridizationisbeyondthescopeofthecurrentstudyitislikelythatthesehybridizationevents have impacted the genetic diversity of eastern coyotesFuturestudiesshouldincluderepresentativeindividualsfromthesepotential introgressing species (redwolves easternwolves graywolvesanddogs)todirectlydeterminetheimpactofhybridizationon the genome‐wide trends of diversity in eastern coyote popu‐lations Interestinglywe note that if hybridization is responsiblefor the observed heterozygosity trends our results suggest thatinterspecific hybridization is most prevalent in the southeasternexpansion frontwhichhas receivedcomparatively lessattentionthan coyotewolf hybridization in the northeastern expansionfront especially on a genome‐wide scale Accordingly redwolfcoyotehybridizationparticularlyoutsideoftheredwolfrecoveryareainNorthCarolina(ieearlyrangeexpansionhybridization)isanintriguingareaforfutureresearch

Our results with respect to genomic diversity are similarto those of vonHoldt et al (2011) who observed comparablepatternsofheterozygosityacrosssimilargroupsofeasterncoy‐ote populationsHowever our observed heterozygosity values(AverageHE =0028)areapproximatelyoneorderofmagnitudelower than those reported by vonHoldt et al (2011) (AverageHE=022)Whilebothestimatesarebasedongenome‐wideSNPdata thisdiscrepancy is likely reflectiveof themethodologicaldifferencesoftheSNPascertainmentstrategiesThatistheca‐nine genotyping array employed in vonHoldt et al (2011) tar‐geted genomic regions that had been previously screened fordiversitywhereas theRADseqmethods used in this study areSNP discovery pipeline without a priori information regardingdiversity

Of the twelveSNPswe identifiedasoutliers several are lo‐catedwithinorneargenes thathavebeen implicated inpheno‐typic traitsnamelydispersalbehaviors thatmaybe relevant torange expansion The behavioral consequences of reduced orcompletely inhibitedgenefunctionatthree loci (CACNA1CALKandEPHA6)wereinvestigatedextensivelyinarodentmodelForexamplemiceheterozygous for aCACNA1C knockoutexhibitedreduced locomotionburstsandscanningbehavioraswellas in‐creased freezing time relative to their wild type counterparts(Kabitzke et al 2017) AdditionallyALK knockout mice exhibitenhancedperformanceinnovelobjectrecognitiontestssuggest‐ingthatthisgeneplaysaroleintheabilityofanimalstoexploreanovelenvironment(Bilslandetal2008)FinallyEPHA6knock‐outmicedemonstratedlearningandspatialmemorydeficitsrel‐ativetowildtypemiceThesetraitsmayallbe intuitively linkedtomovement and dispersal capabilitieswhich are strongly tiedTA

BLE

2emspOutlierSNPsputativelyassociatedwithrangeexpansionidentifiedinboththeoutlierdetectionanalyses

Chr

Posi

tion

Gen

eG

ene

desc

riptio

nG

ene

Regi

onBF

Nor

thea

stBF

Sou

thea

stp

Nor

thea

stp

Sout

heas

t

243210227

5S_r

RNA

5SribosomalRNA

Prom

53564

5477400

13times10

minus597times10

minus7

1015795366

KCN

C2Potassiumvoltage‐gatedchannelsubfamilyCmember2

Intron

3891

130670000

15times10

minus516times10

minus7

1153204490

PAX5

Pairedbox5

Intron

2533

209

23times10

minus318times10

minus5

1653466454

WD

R17

WDrepeatdomain17

Intron

275

5244

35times10

minus326times10

minus6

1723407156

ALK

ALKreceptortyrosinekinase

Intron

499

1050300

11times10

minus415times10

minus13

203062963

EFCC

1EF‐handandcoiled‐coildomaincontaining1

Intron

26286

64928

65times10

minus672times10

minus5

2040628561

ATRI

PATRinteractingprotein

Intron

122

24743

66times10

minus718times10

minus4

2744159565

CACN

A1C

Calciumvoltage‐gatedchannelsubunitalpha1C

Prom

2054

1534times10

minus312times10

minus4

2810746279

ZDH

HC1

6ZincfingerDHHC‐typecontaining16

Intron

178

56046600

18times10

minus349times10

minus5

334379522

EPH

A6EPHreceptorA6

Intron

7073

1378

13times10

minus341times10

minus8

3411841558

AHRR

Aryl‐hydrocarbonreceptorrepressor

Intron

653

155940

33times10

minus452times10

minus8

3523379157

CARM

IL1

Cappingproteinregulatorandmyosin1linker1

Intron

2325

2316

11times10

minus459times10

minus7

Not

esChrChromosomeBFBayesFactorprompromoter

FullBiologicalProcessGOannotationsforeachoutlieraregiveninSupportinginformationTableS6

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 11: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp11HEPPENHEIMER Et al

tospatiallearning(DelgadoPenterianiNamsampCampioni2009Saastamoinenetal2017)

Whilethesefindsareintriguingtheimplicationsforhowmu‐tationsintheseoutlierlociimpactdispersalcapabilitiesincoyotesshouldbeinterpretedwithextremecautionNoneoftheseoutlierSNPswerefoundwithinanexonandit isunclearfromthisdatawhetherthegenotypedvariantsdirectlyimpactgeneexpressionorwhetherthesevariantsareinlinkagedisequilibriumwithoneormore nonsynonymousmutations in coding regions Further evi‐dence for a dispersal‐related function of these genes is entirelybasedonmousestudiesanditisunknownifthesefunctionsareconservedacrossmammalsWealsodidnot incorporateanybe‐havioral data for coyotes in this study and it is unclear if east‐ern coyotes exhibit differences in exploratory behavior relativetocoyotesfromthehistoricalrangemuchlessifthisbehavioriscorrelatedwithgenotypeFuturestudiesshouldtargetthecodingsequenceofthesegenesincoyotestofurtherelucidatehowtheseoutlier locimay influencephenotypic traits in expanding coyotepopulations

It is additionally possible that our outlier detection approachidentified loci that are systematically different between both ex‐pansion fronts and the historical range as a result of parallel ad‐aptationstosimilarenvironmentsratherthantherangeexpansionprocess While the northern and southern expansion fronts aredistinctecoregions(OmernikampGriffith2014)environmentalsimi‐laritiesbetweentheregionsdoexistmostnotablyinthehighabun‐danceofdeerHoweverdietstudiesrevealthatdeerconsumptionvarieswidely amongeastern coyotepopulations (KilgoRayRuthampMiller2010Mastro2011RobinsonDiefenbachFullerHurstampRosenberry2014)andthatdeerconsumptionisalsoreasonablycommon throughout the historical range (Ballard Lutz KeeganCarpenterampdeVosJr2001Carreraetal2008GeseampGrothe1995)Assuchselectionassociatedwiththerangeexpansionpro‐cess is perhapsmore likely than adaptation to deer rich environ‐mentsthoughthepossibilityremainsthatoutlierSNPfrequencies

aredrivenbyselectionassociatedwithunmeasuredenvironmentalvariablesratherthanbyrangeexpansion

OneadditionaloutlierlocusZDHHC16whichhasputativefunc‐tionsrelatedtoeyedevelopmentcellularresponsetoDNAdamageheartdevelopmentandproteinpalmitoylationwasmonomorphicinthehistoricalpopulationyetpolymorphicinbothrecentlyexpandedpopulationsTherearethreegeneralexplanationsfortheoriginofthisvariantintherecentlyexpandedeasterncoyotepopulations(a)Themutationwaspresentinthehistoricalrangebutatanextremelylowfrequencythatwasnotcapturedbyoursampling(b)adenovomutation occurred either convergently along both expansionsfrontsoralongonefrontandthenwastransferredviaintraspecificgeneflowor(c)thisalleleintrogressedfromacloselyrelatedCanis speciesfollowingahybridizationeventandwassubsequentlytrans‐ferredviageneflowAsdiscussedaboveinterspecieshybridizationoccurs among Canis species and there is evidence that this hasbeenadaptiveinthecontextofcoyoterangeexpansion(ThorntonampMurray2014vonHoldtetal2016) It is thereforeconceivablethatZDHHC16representsanadditionalcaseofadaptiveintrogres‐sionHoweverthedatapresentedherearenotsufficienttoaddressquestionsrelatedtotheoriginofgenomicvariantsthoughthisre‐mainsaninterestingquestionforfuturestudiesregardingtheroleofhybridizationinfacilitatingrangeexpansion

Taken together we present a comprehensive genome‐widesurveyof coyotepopulations acrossmuchof the contiguousUSaswellassoutheasternCanadaDespitepronouncedgeographicstructuringamongthehistoricalandtworecentlyexpandedeast‐ern coyote populations we did not observe a strong decline ingenomicdiversitythatischaracteristicofarangeexpansionbottle‐necksuggestingthatcoyoterangeexpansiondynamicsaremorecomplexthanthosedescribedintheoretical(Excoffieretal2009)andotherempiricalstudies(egWhiteetal2013Swaegersetal2015) and is likelyattributable to interspecieshybridizationFurtherweidentifyseveralgenomicvariantsthatarecandidatesfor gene regions under selection during range expansionwhich

F I G U R E 5 emsp (a)OutlierSNPcountsidentifiedbytheBAYENV2andPCadaptanalyses(b)ChangeinallelefrequenciesoftheoutlierSNPsidentifiedinbothanalysesinthenortheastandsoutheastexpansionfrontsrelativetothehistoricalrange

PCadapt

Southeast amp historical

Southeast amp historical

Bayenv2 1253 22

250

231 187

Northeast amp historical

Northeast amp historical

146

141

70

Overlapping oulier genic SNPs(a)

5S rRNA KCNC2 PAX5 WDR17 ALK EFCC1 ATRIP CACNA1C ZDHHC16 EPHA6 AHRR CARMIL1

Outlier SNP

Δ A

llele

freq

uenc

y

minus60

minus40

minus20

0

20

40

60 Southeastern expansion

Northeastern expansion(b)

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 12: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

12emsp |emsp emspensp HEPPENHEIMER Et al

providesacritical firststep inunderstandinghowfunctionalge‐nomicvariationmayhavefacilitatedcoyoterangeexpansion

ACKNOWLEDG EMENTS

WethankAScottenASonnekBConantCWilsonDFrydaDHansenDCaudill JBennetJKittsickJCDaviesJEShermanKQuevieauxKMcGrath LPalmer LPeeblesMEarthmanMDummoldMAndersonMMcKnightPFlicekTQuinnTSullivanK VanWhy R KWayne as well as many hunters and trappersfor sample donation We are also grateful to Oklahoma Museumof Natural History who provided samples under loan agreement

G201634318WethankJohnBensonforfeedbackonanearlierversion of this manuscript This material is based uponwork sup‐ported by the National Science Foundation Graduate ResearchFellowshipunderGrantNoDGE1656466andtheNationalScienceFoundationPostdoctoralResearchFellowshipinBiologyunderGrantNo1523859Thefindingsandconclusionsinthisarticlearethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheUSFishandWildlifeService

CONFLIC T OF INTERE S T

Theauthorsdeclarenoconflictofinterest

F I G U R E 6 emspAllelefrequenciesforalltwelveoutlierSNPsacrosssamplinglocationsAlleleonewasarbitrarilydefinedasthemostcommonalleleoverall(iethemajorallele)

Historical range (pre-1900) Frequency of allele oneFrequency of allele two

EPHA6 AHRR CARMIL1

ATRIP CACNA1C ZDHHC16

ALK WDR17 EFCC1

5S rRNA KCNC2 PAX5

Recently expanded range

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 13: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp13HEPPENHEIMER Et al

AUTHOR CONTRIBUTIONS

EHKEBandBMVdesignedthestudyEHKEBALDandLYRper‐formedDNAextractionsandpreparedlibrariesforsequencingEHperformedanalysesanddraftedthemanuscriptPAHprovidedtech‐nical guidanceon laboratorywork andbioinformaticsKEB JWHBRPTWSRFRKBNWandMJCdonatedsamplesAllauthorscon‐tributedtoandapprovedthefinalmanuscript

DATA ACCE SSIBILIT Y

RADseq clone‐filtered demultiplexed fastq files and CanFam31mapped bam files have been deposited in the NCBI SequenceRead Archive at accession PRJNA497119 (SAMN10248298‐SAMN10248691)Genotypecalls for394coyotesat22935SNPsand associated sample metadata have been deposited to Dryadhttpsdoi105061dryad7f9q2cd

ORCID

Elizabeth Heppenheimer httpsorcidorg0000‐0002‐9164‐3436

Joseph W Hinton httpsorcidorg0000‐0002‐2602‐0400

Linda Y Rutledge httpsorcidorg0000‐0002‐6008‐2322

Alexandra L DeCandia httpsorcidorg0000‐0001‐8485‐5556

R E FE R E N C E S

Abraham G amp Inouye M (2014) Fast principal component analysisof large‐scale genome‐wide data PLoS One 9(4) 1ndash5 httpsdoiorg101371journalpone0093766

Adams J R Leonard J AampWaits L P (2003)Widespread occur‐rence of a domestic dog mitochondrial DNA haplotype in south‐easternUScoyotesMolecular Ecology12(2) 541ndash546httpsdoiorg101046j1365‐294X200301708x

AlexanderDHNovembre Jamp LangeK (2009) Fastmodel‐basedestimationofancestryinunrelatedindividualsGenome Research191655ndash1664httpsdoiorg101101gr094052109

AliOAOrsquoRourkeSMAmishSJMeekMHLuikartGJeffresCampMillerMR(2016)Radcapture(rapture)Flexibleandefficientsequence‐basedgenotypingGenetics202(2)389ndash400httpsdoiorg101534genetics115183665

ArianiABernyMieryTeranJCampGeptsP(2017)Spatialandtemporalscalesofrangeexpansion inwildPhaseolus vulgaris Molecular Biology and Evolution35(1)119ndash131httpsdoiorg101093molbevmsx273

Balestrieri A Remonti L Ruiz‐Gonzaacutelez A Goacutemez‐Moliner B JVergaraMampPrigioniC(2010)Rangeexpansionofthepinemar‐ten (Martes martes) in an agricultural landscapematrix (NW Italy)Mammalian Biology 75(5) 412ndash419 httpsdoiorg101016jmambio200905003

BallardWBLutzDKeeganTWCarpenterLHampdeVos JC(2001) Deer‐predator relationships A review of recent NorthAmerican studies with emphasis on mule and black‐tailed deerWildlife Society Bulletin2999ndash115

BassAJDabneyAampRobinsonD(2018)qvalue Q‐value estimation for false discovery rate control R package version 2140 Retrievedfromhttpgithubcomjdstoreyqvalue

Bilsland J G Wheeldon A Mead A Znamenskiy P AlmondS Waters K A hellip Munoz‐Sanjuan I (2008) Behavioral and

neurochemicalalterationsinmicedeficientinanaplasticlymphomakinase suggest therapeutic potential for psychiatric indicationsNeuropsychopharmacology33(3)685ndash700httpsdoiorg101038sjnpp1301446

Bohling JHDellinger JMcVey JMCobbDTMoormanCEampWaitsLP(2016)DescribingadevelopinghybridzonebetweenredwolvesandcoyotesineasternNorthCarolinaUSAEvolutionary Applications9(6)791ndash804httpsdoiorg101111eva12388

Braschler B amp Hill J K (2007) Role of larval host plants in theclimate‐driven range expansion of the butterfly Polygonia c‐album Journal of Animal Ecology 76(3) 415ndash423 httpsdoiorg101111j1365‐2656200701217x

BurtonOJPhillipsBLampTravisJMJ(2010)Trade‐offsandtheevo‐lutionoflife‐historiesduringrangeexpansionEcology Letters13(10)1210ndash1220httpsdoiorg101111j1461‐0248201001505x

CarbonS IrelandAMungallCJShuSQMarshallBampLewisS (2009) AmiGO Online access to ontology and annotationdata Bioinformatics 25(2) 288ndash289 httpsdoiorg101093bioinformaticsbtn615

CarreraRBallardWGipsonPKellyBTKrausmanPRWallaceMChellipWebsterDB(2008)ComparisonofMexicanwoldandcoy‐otedietsinArizonaandNewMexicoJournal of Wildlife Managament72(2)376ndash381

Catchen J Hohenlohe P A Bassham S Amores A amp CreskoWA (2013) Stacks An analysis tool set for population genomicsMolecular Ecology 22(11) 3124ndash3140 httpsdoiorg101111mec12354

Colautti R I amp Barrett S C H (2013) Rapid adaptation to climatefacilitatesrangeexpansionofan invasiveplantScience342(6156)364ndash366httpsdoiorg101126science1242121

CoopGWitonskyD Rienzo AD amp Pritchard J K (2010)Usingenvironmental correlations to identify loci underlying local ad‐aptation Genetics 185(4) 1411ndash1423 httpsdoiorg101534genetics110114819

RCoreTeam(2013)R A language and environment for statistical comput‐ingViennaAustriaRFoundationforStatisticalComputinghttpswwwR‐projectorg

DeBowTWebsterWampSumner P (1998) Rangeexpansionof thecoyoteCanis latrans (CarnivoraCanidae)intoNorthCarolinawithcomments on some management implications The Journal of the Elisha Mitchell Scientific Society114(3)113ndash118

Delgado M M Penteriani V Nams V O amp Campioni L (2009)Changes of movement patterns from early dispersal to settle‐mentBehavioral Ecology and Sociobiology64(1)35ndash43httpsdoiorg101007s00265‐009‐0815‐5

DraySampDufourAB (2007)Theade4Package Implementing thedualitydiagram for ecologists Journal of Statistical Software22(4)1ndash20httpsdoiorg1018637jssv022i04

Edmonds C A Lillie A S amp Cavalli‐Sforza L L (2004) Mutationsarising in the wave front of an expanding population Proceedings of the National Academy of Sciences 101(4) 975ndash979 httpsdoiorg101073pnas0308064100

Excoffier L Foll M amp Petit R J (2009) Genetic consequencesof range expansions Annual Review of Ecology Evolution and Systematics 40(1) 481ndash501 httpsdoiorg101146annurevecolsys39110707173414

GeseEMampGrotheS(1995)AnalysisofcoyotepredationondeerandelkduringwinterinYellowstoneNationalParkWyomingAmerican Midland Naturalist13336ndash43

GralkaMStieweFFarrellFMoumlbiusWWaclawBampHallatschekO(2016)Allelesurfingpromotesmicrobialadaptationfromstand‐ingvariationEcology Letters19889ndash898httpsdoiorg101111ele12625

Greenbaum G Templeton A R Zarmi Y amp Bar‐David S (2014)Allelicrichnessfollowingpopulationfoundingevents‐Astochastic

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 14: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

14emsp |emsp emspensp HEPPENHEIMER Et al

modelingframeworkincorporatinggeneflowandgeneticdriftPLoS One9(12)1ndash23httpsdoiorg101371journalpone0115203

Guumlnther T amp Coop G (2013) Robust identification of local adapta‐tionfromallele frequenciesGenetics195(1)205ndash220httpsdoiorg101534genetics113152462

HagenSBKopatzAAspiJKojolaIampEikenHG(2015)Evidenceofrapidchange ingeneticstructureanddiversityduringrangeex‐pansioninarecoveringlargeterrestrialcarnivoreProceedings of the Royal Society B Biological Sciences282(1807)20150092httpsdoiorg101098rspb20150092

HeppenheimerECosioDSBrzeskiKECaudillDVanWhyKChamberlainMJhellipvonHoldtB(2018)Demographichistoryin‐fluences spatial patterns of genetic diversityin recently expandedcoyote(Canis latrans)populationsHeredity120(3)183ndash195httpsdoiorg101038s41437‐017‐0014‐5

HillJKCollinghamYCThomasCDBlakeleyDSFoxRMossD amp Huntley B (2001) Impacts of landscape structure on but‐terfly range expansion Ecology Letters 4(4) 313ndash321 httpsdoiorg101046j1461‐0248200100222x

Hinton JWGittleman J L vanManen F TampChamberlainM J(2018) Size‐assortative choice andmate availability influenceshy‐bridizationbetween redwolves (Canis rufus) andcoyotes (Canis la‐trans) Ecology and Evolution83927ndash3940httpsdoiorg101002ece33950

HodyJWampKaysR(2018)Mappingtheexpansionofcoyotes(Canis latrans)acrossAmericaZooKeys9781ndash97httpsdoiorg103897zookeys75915149

HoferTRayNWegmannDampExcoffierL(2009)Largeallelefre‐quency differences between human continental groups are morelikely to have occurred by drift during range expansions than byselection Annals of Human Genetics 73(1) 95ndash108 httpsdoiorg101111j1469‐1809200800489x

Hughes C L Dytham C amp Hill J K (2007) Modelling and ana‐lysing evolution of dispersal in populations at expanding rangeboundaries Ecological Entomology 32(5) 437ndash445 httpsdoiorg101111j1365‐2311200700890x

IbrahimKMNicholsRAampHewittGM (1996)Spatialpatternsofgeneticvariationgeneratedbydifferentformsofdispersalduringrange expansion Heredity 77 282ndash291 httpsdoiorg101038hdy1996142

KabitzkePABrunnerDHeDFazioPACoxKSutphenJhellipClaytonAL(2017)ComprehensiveanalysisoftwoShank3andtheCacna1cmousemodels of autism spectrum disorderGenes Brain and Behavior174ndash22httpsdoiorg101111gbb12405

KaysRCurtisAampKirchmanJJ(2010)RapidadaptiveevolutionofnortheasterncoyotesviahybridizationwithwolvesBiology Letters6(1)89ndash93httpsdoiorg101098rsbl20090575

KilgoJCRayHSRuthCampMillerKV(2010)Cancoyotesaffectdeerpopulations insoutheasternNorthAmericaThe Journal of Wildlife Management74(5)929ndash933httpsdoiorg1021932009‐263

KlopfsteinSCurratMampExcoffierL (2006)Thefateofmutationssurfing on the wave of a range expansionMolecular Biology and Evolution23(3)482ndash490httpsdoiorg101093molbevmsj057

LiHHandsakerBWysokerAFennellTRuanJHomerNMarthGAbecasisGampDurbinR(2009)TheSequencealignmentmapformatandSAMtoolsBioinformatics25(16)2078ndash2079httpsdoiorg101093bioinformaticsbtp352

Lindblad‐TohKWadeCMMikkelsenTSKarlssonEKJaffeDBKamalM LanderES (2005)Genomesequencecompara‐tiveanalysisandhaplotypestructureof thedomesticdogNature438(7069)803ndash819httpsdoiorg101038nature04338

LischerHELampExcoffierL (2012)PGDSpiderAnautomateddataconversion tool for connecting population genetics and genomicsprograms Bioinformatics 28(2) 298ndash299 httpsdoiorg101093bioinformaticsbtr642

Livezey K B (2009) Range expansion of barred owls parts I andII The American Midland Naturalist 161(1) 49ndash56 httpsdoiorg1016740003‐0031‐1612323

LotterhosKEampWhitlockMC(2014)Evaluationofdemographichis‐toryandneutralparameterizationontheperformanceofFSToutliertestsMolecular Ecology23(9)2178ndash2192httpsdoiorg101111mec12725

LotterhosKEampWhitlockMC(2015)Therelativepowerofgenomescans to detect local adaptation depends on sampling design andstatisticalmethodMolecular Ecology24(5)1031ndash1046httpsdoiorg101111mec13100

LunterGampGoodsonM(2011)StampyAstatisticalalgorithmforsen‐sitiveandfastmappingofIlluminasequencereadsGenome Research21(6)936ndash939httpsdoiorg101101gr111120110

LuuKBazinEampBlumMGB (2017)pcadapt AnRpackage toperform genome scans for selection based on principal compo‐nentanalysisMolecular Ecology Resources17(1)67ndash77httpsdoiorg1011111755‐099812592

Mastro L L (2011) Life history and ecology of coyotes in theMid‐AtlanticstatesAsummaryofthescientific literatureSoutheastern Naturalist10(4)721ndash730httpsdoiorg1016560580100411

Mayr E (1954) Change of genetic environment and evolution In JHuxleyACHardyampEB Ford (Eds)Evolution as a process (pp157ndash180)LondonUKAllenampUnwin

McLarenWGilLHuntSERiatHSRitchieGRSThormannA hellip Cunningham F (2016) The Ensembl variant effect pre‐dictor Genome Biology 17(1) 1ndash14 httpsdoiorg101186s13059‐016‐0974‐4

MonzotildenJKaysRampDykhuizenDE(2014)Assessmentofcoyote‐wolf‐dogadmixtureusingancestry‐informativediagnosticSNPsMolecular Ecology23(1)182ndash197httpsdoiorg101111mec12570

NeiMMaruyamaTampChakrabortyR(1975)TheBottleneckeffectandgeneticvariabilityinpopulationsEvolution29(1)1ndash10httpsdoiorg101111j1558‐56461975tb00807x

NoreacutenKStathamMJAringgrenEOIsomursuMFlagstadOslashEideN E hellip Sacks B N (2015) Genetic footprints reveal geographicpatterns of expansion in Fennoscandian red foxesGlobal Change Biology21(9)3299ndash3312httpsdoiorg101111gcb12922

Nowak RM (1979)North American Quaternary Canis Lawrence KSMuseum of Natural History University of Kansas httpsdoiorg105962bhltitle4072

Nowak R M (2002) The original status of wolves in eastern NorthAmerica Southeastern Naturalist 1(2) 95ndash130 httpsdoiorg1016561528‐7092(2002)001[0095TOSOWI]20CO2

Omernik J M amp Griffith G E (2014) Ecoregions of the contermi‐nousUnited States Evolution of a hierarchical spatial frameworkEnvironmental Management541249ndash1266

PatemanRMHill JKRoyDBFoxRampThomasCD (2012)Temperature‐dependentalterationsinhostusedriverapidrangeex‐pansion in a butterflyScience336(6084) 1028ndash1030 httpsdoiorg101126science1216980

Phillips B L Brown G P Travis JM J amp Shine R (2008) ReidrsquosParadox revisited The evolution of dispersal kernels during rangeexpansion The American Naturalist 172(S1) S34ndashS48 httpsdoiorg101086588255

PritchardJKStephensMampDonnellyP (2000) Inferenceofpop‐ulation structure usingmultilocus genotype dataGenetics155(2)945ndash959httpsdoiorg101111j1471‐8286200701758x

Purcell S Neale B Todd‐Brown K Thomas L Ferreira M A RBenderDhellipShamPC (2007)PLINKATool set forwhole‐ge‐nome association and population‐based linkage analyses The American Journal of Human Genetics 81(3) 559ndash575 httpsdoiorg101086519795

RobinsonKFDiefenbachDRFullerAKHurstJEampRosenberryC S (2014) Can managers compensate for coyote predation of

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688

Page 15: High genomic diversity and candidate ... - University of Idaho5Trent University, Peterborough, Ontario 6USFWS National Forensics Laboratory, ... (vonHoldt et al., 2016), it is important

emspensp emsp | emsp15HEPPENHEIMER Et al

white‐tailed deer The Journal of Wildlife Management78(4) 571ndash579httpsdoiorg101002jwmg693

RoussetF (1997)GeneticdifferentiationandestimationofgeneflowfromF‐Statistics under isolation by distanceGenetics145 1219ndash1228httpsdoiorg101002ajmgc30221

RutledgeLYGarrowayCJLovelessKMampPattersonBR(2010)GeneticdifferentiationofeasternwolvesinAlgonquinParkdespitebridging gene flow between coyotes and grey wolves Heredity105(6)520ndash531httpsdoiorg101038hdy20106

SaastamoinenMBocediGCoteJLegrandDGuillaumeFWheatCWhellipdelMarDelgadoM(2017)GeneticsofdispersalBiological Reviews358574ndash599httpsdoiorg101111brv12356

Seehausen O (2004) Hybridization and adaptive radiation Trends in Ecology and Evolution19(4)198ndash207

StreicherJWMcEnteeJPDrzichLCCardDCSchieldDRSmartUhellipCastoeTA(2016)Geneticsurfingnotallopatricdi‐vergence explains spatial sorting of mitochondrial haplotypes invenomous coralsnakes Evolution 70(7) 1435ndash1449 httpsdoiorg101111evo12967

Swaegers JMergeay J Geystelen A V A N amp Therry L (2015)Neutral and adaptive genomic signatures of rapid poleward rangeexpansion Molecular Ecology 24(24) 6163ndash6176 httpsdoiorg101111mec13462

SzpiechZAJakobssonMampRosenbergNA(2008)ADZEArarefac‐tionapproachforcountingallelesprivatetocombinationsofpopu‐lationsBioinformatics24(21)2498ndash2504httpsdoiorg101093bioinformaticsbtn478

TaulmanJFampRobbinsLW(1996)Recentrangeexpansionanddistri‐butionallimitsofthenine‐bandedarmadillo(Dasypus novemcinctus) intheUnitedStatesJournal of Biogeography23(5)635ndash648httpsdoiorg101111j1365‐26991996tb00024x

ThorntonDHampMurrayDL(2014)Influenceofhybridizationonnicheshifts in expanding coyote populationsDiversity and Distributions20(11)1355ndash1364httpsdoiorg101111ddi12253

TravisJMJampDythamC(2002)DispersalevolutionduringinvasionsEvolutionary Ecology Research4(8)1119ndash1129

vonHoldtBHeppenheimerEPetrenkoVCroonquistPampRutledgeLY(2017)Ancestry‐specificmethylationpatternsinadmixedoff‐springfromanexperimentalcoyoteandgrayWolfcrossJournal of Heredity108(4)341ndash348httpsdoiorg101093jheredesx004

vonHoldt B M Kays R Pollinger J P amp Wayne R K (2016)AdmixturemappingidentifiesintrogressedgenomicregionsinNorthAmericancanidsMolecular Ecology25(11)2443ndash2453httpsdoiorg101111mec13667

vonHoldtBMPollingerJPEarlDAKnowlesJCBoykoARParkerHhellipWayneRK(2011)Agenome‐wideperspectiveontheevolutionaryhistoryofenigmaticwolf‐likecanidsGenome Research21(8)1294ndash1305httpsdoiorg101101gr116301110

VossNEcksteinRLampDurkaW(2012)Rangeexpansionofaself‐ingpolyploidplantdespitewidespreadgeneticuniformityAnnals of Botany110(3)585ndash593httpsdoiorg101093aobmcs117

WangJDuncanDShiZampZhangB(2013)WEB‐basedGEneSeTAnaLysisToolkit(WebGestalt)Update2013Nucleic Acids Research4177ndash83httpsdoiorg101093nargkt439

WheeldonTPattersonBampWhiteB(2010)ColonizationhistoryandancestryofnortheasterncoyotesBiology Letters6(2)246ndash247au‐thorreply248ndash249httpsdoiorg101098rsbl20090822

WheeldonTJRutledgeLYPattersonBRWhiteBNampWilsonP J (2013) Y‐chromosomeevidence supports asymmetric dog in‐trogressionintoeasterncoyotesEcology and Evolution3(9)3005ndash3020httpsdoiorg101002ece3693

WhiteTAPerkinsSEHeckelGampSearle JB (2013)AdaptiveevolutionduringanongoingrangeexpansionTheinvasivebankvole(Myodes glareolus) in IrelandMolecular Ecology22(11) 2971ndash2985httpsdoiorg101111mec12343

WilsonPJRutledgeLYWheeldonTJPattersonBRampWhiteBN (2012) Y‐chromosome evidence supportswidespread signa‐turesofthree‐speciescanishybridizationineasternNorthAmericaEcology and Evolution 2(9) 2325ndash2332 httpsdoiorg101002ece3301

YoungSPampJacksonHHT(1951)The clever coyoteHarrisburgPATheStackpoleCompany

ZerbinoDRAchuthanPAkanniWAmodeMRBarrellDBhaiJhellipFlicekP(2017)Ensembl2018Nucleic Acids Research46754ndash761httpsdoiorg101093nargkx1098

ZhangBKirovSampSnoddyJ(2005)WebGestaltAnintegratedsys‐tem for exploring gene sets in various biological contextsNucleic Acids Research33741ndash748httpsdoiorg101093nargki475

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupportingInformationsectionattheendofthearticle

How to cite this articleHeppenheimerEBrzeskiKEHintonJWetalHighgenomicdiversityandcandidategenesunderselectionassociatedwithrangeexpansionineasterncoyote(Canis latrans)populationsEcol Evol 2018001ndash15 httpsdoiorg101002ece34688