Aalborg Universitet

Disentangling Timing of Admixture Patterns of Introgression and PhenotypicIndicators in a Hybridizing Wolf Population

Galaverni Marco Caniglia Romolo Pagani Luca Fabbri Elena Boattini Alessio RandiEttorePublished inMolecular Biology and Evolution

DOI (link to publication from Publisher)101093molbevmsx169

Creative Commons LicenseCC BY-NC 40

Publication date2017

Document VersionPublishers PDF also known as Version of record

Link to publication from Aalborg University

Citation for published version (APA)Galaverni M Caniglia R Pagani L Fabbri E Boattini A amp Randi E (2017) Disentangling Timing ofAdmixture Patterns of Introgression and Phenotypic Indicators in a Hybridizing Wolf Population MolecularBiology and Evolution 34(9) 2324-2339 httpsdoiorg101093molbevmsx169

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors andor other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights

Users may download and print one copy of any publication from the public portal for the purpose of private study or research You may not further distribute the material or use it for any profit-making activity or commercial gain You may freely distribute the URL identifying the publication in the public portal

Take down policyIf you believe that this document breaches copyright please contact us at vbnaubaaudk providing details and we will remove access tothe work immediately and investigate your claim

Disentangling Timing of Admixture Patterns of Introgressionand Phenotypic Indicators in a Hybridizing Wolf Population

Marco Galaverni12 Romolo Caniglia1 Luca Pagani34 Elena Fabbri1 Alessio Boattini5 and Ettore Randi16

1Area per la Genetica della Conservazione ISPRA Ozzano dellrsquoEmilia Bologna Italy2Area Conservazione WWF Italia Rome Italy3Dipartimento di Biologia Universita degli Studi di Padova Padua Italy4Estonian Biocentre Tartu Estonia5Department of Biological Geological and Environmental Sciences University of Bologna Bologna Italy6Department 18Section of Environmental Engineering Aalborg Universitet Aalborg Denmark

Corresponding authors E-mails marcogalaverniisprambienteit mgalaverniwwfit

Associate editor Daniel Falush

Abstract

Hybridization is a natural or anthropogenic process that can deeply affect the genetic make-up of populations possiblydecreasing individual fitness but sometimes favoring local adaptations The population of Italian wolves (Canis lupus)after protracted demographic declines and isolation is currently expanding in anthropic areas with documented cases ofhybridization with stray domestic dogs However identifying admixture patterns in deeply introgressed populations is farfrom trivial In this study we used a panel of 170000 SNPs analyzed with multivariate Bayesian and local ancestryreconstruction methods to identify hybrids estimate their ancestry proportions and timing since admixture Moreoverwe carried out preliminary genotypendashphenotype association analyses to identify the genetic bases of three phenotypictraits (black coat white claws and spur on the hind legs) putative indicators of hybridization Results showed no sharpsubdivisions between nonadmixed wolves and hybrids indicating that recurrent hybridization and deep introgressionmight have started mostly at the beginning of the population reexpansion In hybrids we identified a number of genomicregions with excess of ancestry in one of the parental populations and regions with excess or resistance to introgressioncompared with neutral expectations The three morphological traits showed significant genotypendashphenotype associa-tions with a single genomic region for black coats and white claws and with multiple genomic regions for the spur In allcases the associated haplotypes were likely derived from dogs In conclusion we show that the use of multiple genome-wide ancestry reconstructions allows clarifying the admixture dynamics even in highly introgressed populations andsupports their conservation management

Key words admixture history Canis lupus genotypendashphenotype association hybridization introgression

IntroductionHybridization is a biological process that through the cross-breeding between individuals from distinct but closely relatedtaxa or between discrete entities that exchange genes candeeply affect their genetic make-up long-term survival andevolution (Parham et al 2013 Saetre 2013 Gompert andBuerkle 2016) Natural hybridization is no longer viewed asa sporadic and undesirable evolutionary dead-end but ratheras a relatively frequent and potentially creative process(Mallet 2008 Larsen et al 2010 Bailey et al 2013 Stern2013 Rius and Darling 2014) Natural hybrid zones representhot-spots of genetic diversity where natural selection canexpose novel gene assemblages to the adaptive process pos-sibly leading to hybrid speciation (Lavrenchenko andBulatova 2016) Episodic hybridization may introduce geneticvariation into isolated populations contrasting the deleteri-ous consequences of small effective size and inbreeding (ge-netic rescue Brennan et al 2015) and cases of adaptiveintrogression have been recently documented in diverse

animal and plant species (Grossen et al 2014 Hovick andWhitney 2014 Fulgione et al 2016) Conversely anthropo-genic hybridization caused by habitat changes and by thespread of alien or domestic species is mostly viewed as arisk factor in conservation biology especially for the deleteri-ous consequences of gene introgression into wild populationsthrough backcrossing (Allendorf et al 2001 Laikre et al 2010)Introgression may spread maladaptive variants causing fitnessdeclines and outbreeding depression swamp genetic diver-sity or disrupt epistatic equilibria and local adaptations driv-ing local populations or entire species to extinction (genomicextinctions Rhymer and Simberloff 1996 Todesco et al 2016)However the long-term evolutionary consequences of ge-netic admixtures remain largely unpredictable and can some-times be beneficial (Coulson et al 2011 Hedrick et al 2016)

Hybridized populations may have complex historiesHybrids can be generated either by discrete events (egfrom single migrations waves) or by recurrent gene flow

Article

The Author 2017 Published by Oxford University Press on behalf of the Society for Molecular Biology and EvolutionThis is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License(httpcreativecommonsorglicensesby-nc40) which permits non-commercial re-use distribution and reproduction in anymedium provided the original work is properly cited For commercial re-use please contact journalspermissionsoupcom Open Access2324 Mol Biol Evol 34(9)2324ndash2339 doi101093molbevmsx169 Advance Access publication May 26 2017Downloaded from httpsacademicoupcommbearticle-abstract34923243855352

by Aalborg University Library useron 23 April 2018

and admixture (eg in stable areas of sympatry Allendorfet al 2001 Currat et al 2008 Moorjani et al 2011 Muhlfeldet al 2014) Whereas modeling the former case is relativelysimple since the relative proportions of ancestral compo-nents will be fairly homogeneous across all the hybrid indi-viduals (Rhymer and Simberloff 1996 Rieseberg et al 1999Rosenberg et al 2010) identifying the patterns and timing ofrecurrent hybridization events is far more complex (Curratand Excoffier 2011 Twyford and Ennos 2012) and is condi-tioned by the availability of efficient ancestry-informativemarkers (AIMs) Moreover hybridization is expected to affectthe phenotypes of admixed individuals depending on thedominance and quantitative nature of the considered traits(Mallet 2008 Patterson et al 2010) Therefore the study ofhybridization and introgression in natural populations alsoprovides the opportunity to identify loci associated with phe-notypic features adaptation and selection which might un-derlie the genetic bases of fitness in hybrid individuals (Arnoldand Martin 2010 Gompert and Buerkle 2012)

Canids (genus Canis) represent interesting examples ofhybridizing taxa sharing related and recently divergent ge-nomes In Eurasia wolves (C lupus) can hybridize in the wildwith the closely related golden jackals (C aureus Freedmanet al 2014 Moura et al 2014) In North America historicalwolf x coyote (C latrans) hybridization originated complexpatterns of admixed taxa such as the red wolf (C rufus)although their origin and taxonomic status is highly debated(Fredrickson and Hedrick 2006 Koblmuller et al 2009VonHoldt et al 2011 Gese et al 2015 Rutledge et al 2015VonHoldt et al 2016a 2016b) Wolves coyotes and goldenjackals can successfully reproduce also with domestic dogs(C l familiaris) principally in human-dominated regionswhere the widespread diffusion of stray dogs threatens thegene pool of several wild canid populations constituting aserious conservation concern (Stronen and Paquet 2013Monzon et al 2014 Randi et al 2014 Galov et al 2015)

A particular case-study is offered by the Italian wolf pop-ulation which is sharply genetically differentiated from anyother wolf population due to protracted isolation south ofthe Alps and to demographic declines that led it to shrinkdown to lt100 individuals in the 1970s (Zimen and Boitani1975 Lucchini et al 2004 Pilot et al 2014 Randi et al 2014)Thanks to the increased availability of prey and legal protec-tion Italian wolves are now recovering (Galaverni et al 2016)but they are still threatened by heavy poaching (Caniglia et al2010 Imbert et al 2016) and hybridization with stray dogs asdocumented by genetic evidences (Lucchini et al 2004Caniglia et al 2013 Randi et al 2014)

To date except for a few studies where SNP arrays wereapplied but not to assess hybridization patterns in this pop-ulation (Vonholdt et al 2011 Stronen and Paquet 2013 Pilotet al 2014) Italian wolves have been genotyped using smallpanels of microsatellites (STRs) and uniparental markers (mi-tochondrial DNA and Y-chromosome sequences) which al-low the detection of hybridization only up to 2ndash3 generationsin the past (Veuroaheuroa and Primmer 2006 Randi 2008) but not tounderstand the temporal patterns of older hybridization

events In particular it is unknown whether the Italian wolvesprevalently hybridized during or before the bottleneck of the1970s or if they recurrently admixed during the reexpansionalong the Apennine ridge in the last 40 years To overcomethese limitations in this study we genotyped 118 wolves 31village dogs and 72 putative genetic andor phenotypical wolfx dog hybrids in addition to 456 publicly available dog geno-types at a panel of 170000 genome-wide single nucleotidepolymorphisms (SNPs) and we exploited the additional in-formation inferred from their haplotype blocks aiming toidentify admixed genotypes older than the first few genera-tions of backcrossing and to estimate their timing of admix-ture (Lindblad-Toh et al 2005 Vonholdt et al 2010 Pattersonet al 2012) Furthermore we applied local ancestry recon-struction methods to identify regions of the genomes of hy-brids that were significantly deviating from randominheritance patterns likely indicating selective forces locallyacting on such regions Moreover wolf x dog hybridizationmay be reflected at the phenotypic level The causative mu-tations of some phenotypic variants are well known such asthe one for the black coat color (Candille et al 2007) Thistrait is coded by a 3-bp deletion at the b-defensin geneCDB103 that was likely introduced into North Americanwolves by ancient hybridization with dogs (Anderson et al2009) Other traits have also been suggested as potential in-dicators of hybridization such as white claws (compared withthe usual black or dark-grey wild phenotype) or the spur onthe hind legs (a relict fifth finger also known as dewclaw orpreaxial polydactyly Deng et al 2015) but their genetic de-terminants are unknown or uncertain (Ciucci et al 2003Caniglia et al 2013 Randi et al 2014) Therefore we usedlocal genome ancestry (Gompert and Buerkle 2012) and ge-notypendashphenotype association procedures (Gorlova et al2011) to identify putative dog-derived causal mutations as-sociated with phenotypic variants that might be introgressedinto the Italian wolves

Results

Data Filtering Sample and Marker SelectionFollowing genotyping and quality cleaning steps performed inSVS both per sample and per locus we retained 213 unre-lated samples with high call rate (964) that were success-fully genotyped at 156132 autosomal SNPs (90 hereafterreferred to as the ldquo156k data setrdquo) these samples included 114wolves 68 putative hybrids (identified based on previous STRassignments andor anomalous phenotypic traits Randi et al2014) and 31 village dogs which were added to 456 dogsamples from 30 breeds publicly available from the LUPAproject data set created for the genetic mapping of a numberof canine diseases (Lequarre et al 2011 Vaysse et al 2011) Asubset of 25061 SNPs (145) was obtained after LD pruningperformed in PLINK (Purcell et al 2007) at threshold r2frac14 01(the 25 k data set) A smaller set of 19732 SNPs (114) wasobtained after discarding all sites with any missing data (19 kdata set)

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2325Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

and admixture (eg in stable areas of sympatry Allendorfet al 2001 Currat et al 2008 Moorjani et al 2011 Muhlfeldet al 2014) Whereas modeling the former case is relativelysimple since the relative proportions of ancestral compo-nents will be fairly homogeneous across all the hybrid indi-viduals (Rhymer and Simberloff 1996 Rieseberg et al 1999Rosenberg et al 2010) identifying the patterns and timing ofrecurrent hybridization events is far more complex (Curratand Excoffier 2011 Twyford and Ennos 2012) and is condi-tioned by the availability of efficient ancestry-informativemarkers (AIMs) Moreover hybridization is expected to affectthe phenotypes of admixed individuals depending on thedominance and quantitative nature of the considered traits(Mallet 2008 Patterson et al 2010) Therefore the study ofhybridization and introgression in natural populations alsoprovides the opportunity to identify loci associated with phe-notypic features adaptation and selection which might un-derlie the genetic bases of fitness in hybrid individuals (Arnoldand Martin 2010 Gompert and Buerkle 2012)

Canids (genus Canis) represent interesting examples ofhybridizing taxa sharing related and recently divergent ge-nomes In Eurasia wolves (C lupus) can hybridize in the wildwith the closely related golden jackals (C aureus Freedmanet al 2014 Moura et al 2014) In North America historicalwolf x coyote (C latrans) hybridization originated complexpatterns of admixed taxa such as the red wolf (C rufus)although their origin and taxonomic status is highly debated(Fredrickson and Hedrick 2006 Koblmuller et al 2009VonHoldt et al 2011 Gese et al 2015 Rutledge et al 2015VonHoldt et al 2016a 2016b) Wolves coyotes and goldenjackals can successfully reproduce also with domestic dogs(C l familiaris) principally in human-dominated regionswhere the widespread diffusion of stray dogs threatens thegene pool of several wild canid populations constituting aserious conservation concern (Stronen and Paquet 2013Monzon et al 2014 Randi et al 2014 Galov et al 2015)

A particular case-study is offered by the Italian wolf pop-ulation which is sharply genetically differentiated from anyother wolf population due to protracted isolation south ofthe Alps and to demographic declines that led it to shrinkdown to lt100 individuals in the 1970s (Zimen and Boitani1975 Lucchini et al 2004 Pilot et al 2014 Randi et al 2014)Thanks to the increased availability of prey and legal protec-tion Italian wolves are now recovering (Galaverni et al 2016)but they are still threatened by heavy poaching (Caniglia et al2010 Imbert et al 2016) and hybridization with stray dogs asdocumented by genetic evidences (Lucchini et al 2004Caniglia et al 2013 Randi et al 2014)

To date except for a few studies where SNP arrays wereapplied but not to assess hybridization patterns in this pop-ulation (Vonholdt et al 2011 Stronen and Paquet 2013 Pilotet al 2014) Italian wolves have been genotyped using smallpanels of microsatellites (STRs) and uniparental markers (mi-tochondrial DNA and Y-chromosome sequences) which al-low the detection of hybridization only up to 2ndash3 generationsin the past (Veuroaheuroa and Primmer 2006 Randi 2008) but not tounderstand the temporal patterns of older hybridization

events In particular it is unknown whether the Italian wolvesprevalently hybridized during or before the bottleneck of the1970s or if they recurrently admixed during the reexpansionalong the Apennine ridge in the last 40 years To overcomethese limitations in this study we genotyped 118 wolves 31village dogs and 72 putative genetic andor phenotypical wolfx dog hybrids in addition to 456 publicly available dog geno-types at a panel of 170000 genome-wide single nucleotidepolymorphisms (SNPs) and we exploited the additional in-formation inferred from their haplotype blocks aiming toidentify admixed genotypes older than the first few genera-tions of backcrossing and to estimate their timing of admix-ture (Lindblad-Toh et al 2005 Vonholdt et al 2010 Pattersonet al 2012) Furthermore we applied local ancestry recon-struction methods to identify regions of the genomes of hy-brids that were significantly deviating from randominheritance patterns likely indicating selective forces locallyacting on such regions Moreover wolf x dog hybridizationmay be reflected at the phenotypic level The causative mu-tations of some phenotypic variants are well known such asthe one for the black coat color (Candille et al 2007) Thistrait is coded by a 3-bp deletion at the b-defensin geneCDB103 that was likely introduced into North Americanwolves by ancient hybridization with dogs (Anderson et al2009) Other traits have also been suggested as potential in-dicators of hybridization such as white claws (compared withthe usual black or dark-grey wild phenotype) or the spur onthe hind legs (a relict fifth finger also known as dewclaw orpreaxial polydactyly Deng et al 2015) but their genetic de-terminants are unknown or uncertain (Ciucci et al 2003Caniglia et al 2013 Randi et al 2014) Therefore we usedlocal genome ancestry (Gompert and Buerkle 2012) and ge-notypendashphenotype association procedures (Gorlova et al2011) to identify putative dog-derived causal mutations as-sociated with phenotypic variants that might be introgressedinto the Italian wolves

Results

Data Filtering Sample and Marker SelectionFollowing genotyping and quality cleaning steps performed inSVS both per sample and per locus we retained 213 unre-lated samples with high call rate (964) that were success-fully genotyped at 156132 autosomal SNPs (90 hereafterreferred to as the ldquo156k data setrdquo) these samples included 114wolves 68 putative hybrids (identified based on previous STRassignments andor anomalous phenotypic traits Randi et al2014) and 31 village dogs which were added to 456 dogsamples from 30 breeds publicly available from the LUPAproject data set created for the genetic mapping of a numberof canine diseases (Lequarre et al 2011 Vaysse et al 2011) Asubset of 25061 SNPs (145) was obtained after LD pruningperformed in PLINK (Purcell et al 2007) at threshold r2frac14 01(the 25 k data set) A smaller set of 19732 SNPs (114) wasobtained after discarding all sites with any missing data (19 kdata set)

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2325Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

Assignment and Admixture AnalysesThe first component of an exploratory PCA performed in SVSusing the 156 k data set explained most of the genetic var-iability (72 fig 1) and clearly differentiated dogs fromwolves The putative genetic hybrids were scattered alongthe first axis between wolves and dogs much closer to theformer except one intermediate sample Conversely the pu-tative hybrids identified only by atypical phenotypes weremostly included within the wolf cluster The second axiswhich explained a 10-time smaller portion of the geneticdiversity differentiated the most divergent dog breed theDobermanns followed by Border Terriers and Weimaranersalong the additional axes (supplementary fig S1Supplementary Material online)

Results from ADMIXTURE (Alexander et al 2009) run with the25k data set to reassign each sample to its population of originwere concordant with the PCA analysis (fig 2) At increasingvalues of K the cross validation error steadily decreased up toKfrac14 34 (supplementary fig S2 Supplementary Material on-line) However the first main decrease in CV error was ob-served at Kfrac14 2 that clearly separated wolves from dogs Thuswe retained Kfrac14 2 to identify the proportions of admixture inwolf x dog hybrids All dogs (DOG nfrac14 487) had an individualassignment value qwlt 015 (fig 2) Similarly based on the

distribution of individual values in the reference wolves (fig2) we selected a conservative threshold qw 0999 to assigngenotypes to the reference Italian wolf cluster (WIT nfrac14 90)and all those with an intermediate assignment value as possi-ble hybrids (HYB nfrac14 92) in the downstream analyses inde-pendently of phenotypic information Since at Kgt 10 somelevel of substructuring was observed within wolves we usedKfrac14 10 to identify the prevalent dog components in the ad-mixed individuals which turned out to correspond to GermanShepherds (fig 2) At Kfrac14 34 corresponding to the optimalnumber of genetic clusters (supplementary fig S2Supplementary Material online) all breeds clearly separatedfrom one another (fig 2) and also from wolves which weresplit into five main groups (centred around the NorthernApennines Eastern CoastmdashAdriatic Western CoastmdashMaremma Central Apennines Southern Apennines) reflect-ing a rough geographical substructure along the ApenninesTHREEPOP results for the f3 test (Pickrell and Pritchard 2012)confirmed a highly significant admixture between wolves anddogs in the HYB samples (z-scorefrac1472064) To better esti-mate the actual admixture proportions in the analyzed hybrids(Falush et al 2016) we applied a PCA-based admixture decon-volution approach implemented in PCADMIX (Brisbin et al2012) which identified 12ndash516 dog blocks in the genome

FIG 1 PC1 versus PC2 results from an exploratory principal component analysis (PCA) computed in SVS on the 156 k SNP data set and includingdogs (yellow dots) putatively nonadmixed wolves (blue squares) black wolves (black squares) genetically admixed wolves based on STR data(pink squares) and wolves with atypical phenotypes (gray squares) The two axes are not to scale in order to better distinguish individuals alongPC2 The shaded rectangle includes a zoom on the wolfhybrid cluster The dog cluster in the bottom-right corner of the figure is composed ofDobermann individuals

Galaverni et al doi101093molbevmsx169 MBE

2326Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

of the hybrid individuals These proportions highly correlatedwith those estimated in ADMIXTURE at Kfrac14 2 (R2frac14 0969Plt 001) but highlighted larger dog components (supple-mentary fig S3 Supplementary Material online)

Timing of AdmixtureThe software ALDER (Loh et al 2013) by exploiting informationderived from the haplotype structure based on recombinationrates and the extent of LD decay among neighboring lociidentified significant admixture in each generation cohort orpool of cohorts of putative hybrids (generations G1 G2 G3ndash4and G5ndash9 from the present all with P valueslt 14106 sup-plementary table S1 Supplementary Material online) althoughit reported inconsistent decay rates (meaning that one of theparental populations could have been not fully sampled Lohet al 2013 supplementary fig S4 Supplementary Material on-line) Hybridization was estimated to have occurred 35ndash50generations before sampling and assuming a wolf generationtime of 3 years (Skoglund et al 2011) this corresponded to four

partially overlapping intervals (one per cohort) in which ad-mixture possibly originated centerd around years 1987 19951999 and 2000 (fig 3)

PCADMIX results that we also used to estimate individualadmixture times indicated that admixture mostly occurred3ndash4 generations before sampling with the oldest events trac-ing back up to 19 generations before sampling (supplemen-tary fig S5 Supplementary Material online) Converted intoyears this indicated that the hybridization events likelystarted in the 1940s followed in the late 1970s but peakedin the late 1990s (fig 3) highly concordant with the resultsfrom ALDER Although our methods could be expected to bemore efficient in tracing more recent hybridization events theestimated timing of admixture through PCADMIX appearedrobust since estimates from different generation cohortswere highly concordant (supplementary fig S5Supplementary Material online) From a geographical pointof view the patterns of admixture showed an initial South-to-North trend followed by more complex dynamics (fig 4

FIG 2 ADMIXTURE results from the 25 k SNP data set at Kfrac14 2 Kfrac14 10 and Kfrac14 34 Kfrac14 2 clearly separates wolves from dogs whereas at Kfrac14 10 it ispossible to identify the prevalent dog component in the admixed wolves namely from German Shepherds (in blue) At Kfrac14 34 all the dog breedsclearly separates from one another and wolves are split into five main groups reflecting a rough geographical structure of the population along theApennines (Northern Apennines in yellow Eastern CoastmdashAdriatic in blue Western CoastmdashMaremma in dark gray Central Apennines in pink andSouthern Apennines in light brown) The intermediate hybrid identified in the PCA is the third individual of the HIT subset DITfrac14 village dogs (31)BeT frac14 Belgian Tervuren (12) Bgl frac14 Beagle (10) BMD frac14 Bernese Mountain Dog (12) BoC frac14 Border Collie (16) BoT frac14 Border Terrier (25) BrS frac14Brittany Spaniel (12) CoSfrac14Cocker Spaniel (14) Dacfrac14Dachshund (12) Dobfrac14Doberman Pinscher (25) EBDfrac14 English Bulldog (13) Elkfrac14 Elkhound(12) EStfrac14 English Setter (12) Eurfrac14 Eurasian (12) Finnish Spitz FSp (12) GoSfrac14Gordon Setter (25) Gryfrac14Golden Retriever (11) GRefrac14Greyhound(14) GShfrac14 German Shepherd (12) GSlfrac14 Greenland Sledge Dog (12) IrWfrac14 Irish Wolfhound (11) JRTfrac14 Jack Russell Terrier (12) LRefrac14 LabradorRetriever (14) NFdfrac14Newfoundland (25) NSDfrac14Nova Scotia Duck Tolling Retriever (23) Rtwfrac14 Rottweiler (12) Scifrac14 Schipperke (25) ShPfrac14 Shar-Pei (11) StPfrac14 Standard Poodle (12) TYofrac14 Yorkshire Terrier (12) Weifrac14Weimaraner (26) PITfrac14wolves with atypical phenotypes (10) BITfrac14 blackwolves (14) HIT frac14 genetically admixed wolves based on STR data (45) WIT frac14 putatively nonadmixed Italian wolves (118)

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2327Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

supplementary file S1 Supplementary Material online) likelyreflecting the range expansion of the species along theApennines (Fabbri et al 2007)

Deviations from Admixture NeutralityIn order to look at deviations from neutrality in the inheri-tance patterns of parental haplotypes in hybrids all blocksanalyzed by PCADMIX in the hybrids were ranked according totheir relative proportion of ldquodogrdquo or ldquowolfrdquo assignment andlabeled as admixture outliers if falling into the top or bottom1 of the genome-wide frequency distribution In this waywe identified 73 regions with high frequency of dog alleles inhybrids (mostly mapping to chromosomes chr3 chr10 chr28and chr32) and 106 blocks with high frequency of wolf alleles(mainly on chr7 chr8 and chr12) likely indicating outlier

regions for excess of ancestry from one of the parental pop-ulations (fig 5) The ldquodog-likerdquo outlier regions included 179annotated protein-coding genes significantly enriched forGene Ontology (GO) categories related to transmembranetransport and a number of Human Phenotype (HP) catego-ries the most significant of which were linked to nasal and earmorphology (supplementary file S2 Supplementary Materialonline) The ldquowolf-likerdquo outlier regions included 235 protein-coding genes significantly enriched for immunity-relatedmetabolic and enzymatic biological processes (BP) and forHP categories related to abnormal ossification (supplemen-tary file S2 Supplementary Material online)

Local deviations from neutrality were also investigatethrough a Bayesian Genomic Cline analysis in BGC(Gompert and Buerkle 2012) which identifies regions with

FIG 3 Top timing since the admixture event for each admixed individual deduced from the empirical distribution of the number of chromosomalswitches inferred from PCADMIX in relation to the individual assignment values (proportion of wolf blocks) The expected distributions atincreasing generations since admixture are indicated by the colored lines Bottom temporal distribution of the admixture events deducedfrom PCADMIX (vertical bars) compared with the time intervals reconstructed by ALDER (horizontal bars) from four generation cohorts or pools ofcohorts with G1 being the last sampled generation (years 2013ndash2015) and so on considering a wolf generation time of 3 years (Skoglund et al2011) Square triangle and diamond symbols represent mean values and vertical sticks represent confidence intervals

Galaverni et al doi101093molbevmsx169 MBE

2328Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

excess of ancestry in one of the parental populations com-pared with neutral expectations (a parameter) as well asthose with excess of or resistance to introgression (szlig param-eter) BGC results for the a parameter indicated that 187SNPs distributed throughout most of the chromosomeshad an excess of dog ancestry (significantly positive valuesof a) and 132 SNPs an excess of wolf ancestry (alt 0) withoverall higher absolute values in the former (fig 5) A signif-icant excess of introgression was observed only in six SNPs onchr19 and chr24 (szliglt 0) whereas nine SNPs on chr17 chr21chr27 and chr35 showed values of szliggt 0 indicating a resis-tance to introgression or steeper genetic clines (fig 5) The50-kb regions surrounding the SNPs with excess of dog an-cestry contained 210 protein-coding genes that were mostlyenriched for a GO biological process related to inflammatoryresponse and a HP category linked to short stature (supple-mentary file S2 Supplementary Material online) Converselythe regions with excess of wolf ancestry contained 156 codinggenes enriched for membrane and protein-related cellularcomponents (CC) and for skin-related HP categories (supple-mentary file S2 Supplementary Material online)

The 50-kb regions surrounding the outlier SNPs for the szligparameter included only eight genes with excess of introgres-sion enriched for GO categories related to telomere process-ing and 18 genes with resistance to introgression enriched forGO categories mainly linked to neural cells functions andglucose transport (supplementary file S2 SupplementaryMaterial online) No SNP was significant for both parameters(supplementary fig S2 Supplementary Material online) al-though a region located on chr24 around 46ndash47 Mb included

FIG 4 Map visualizing the geographical distribution of the wolf x dogadmixture events in Italy through time as reconstructed fromPCADMIX results Locations are plotted where the admixed individualshave been sampled and cannot fully reflect the potential movementsfrom where the first parental hybrids were sampled An animatedversion of the map is available in supplementary file S1Supplementary Material online

FIG 5 Distribution of the introgressed regions along the genomes of admixed individuals Top PCADMIX outlier regions Blue bars indicate regionswith excess of wolf-derived alleles yellow bars indicate excess of dog-derived alleles Center BGC (Bayesian Genomic Cline analysis Gompert andBuerkle 2012) alpha parameter outlier SNPs Values higher than 0 indicate excess of dog alleles values lower than 0 indicate excess of wolf allelesBottom outlier SNPs for the BGC beta parameter Values higher than 0 indicate resistance to introgression lower than 0 an excess of introgressioncompared to random expectations In both cases BGC significant outliers are indicated by blue crosses (top or bottom 1 of the empiricaldistribution of values) and by red dots (95 credibility intervals of 10000 iterations not including 0)

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2329Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

both SNPs that were outlier for szliglt 0 and SNPs either sig-nificantly positive or negative for a indicating a possible ex-cess of introgression in this region for both wolf and dogalleles We found only a partial overlap between the outlierSNPs identified from the BGC a parameter and the blocksof ancestry-excess identified from PCADMIX namely a regionon chr25 for dog excess and a region on chr27 for wolfexcess (supplementary file S2 Supplementary Materialonline)

Ancestry-Informative Marker SelectionTo verify their power to identify wolves dogs and their hy-brids we selected three panels of the most ancestry-informative markers (192 96 and 48 SNPs supplementaryfile S2 Supplementary Material online) from the 25 k dataset (FSTfrac14 016 HO-dogfrac14 02312 HO-wolffrac14 01921) and usedthem to perform additional PCA analyses Results were highlyconcordant and well differentiated dogs and wolves (192 topSNPs FSTfrac14 058 HO-dogfrac14 03051 HO-wolffrac14 00075 Top 96

SNPs FSTfrac14 060 HO-dogfrac14 02957 HO-wolffrac14 00087 Top 48SNPs FSTfrac14 061 HO-dogfrac14 02843 HO-wolffrac14 00095) Most ofthe admixed individuals clustered in an intermediate positionbetween the parental groups although there was still no clearsubdivision between some of the hybrids and the wolves(fig 6) The assignment values from ADMIXTURE run at Kfrac14 2on the 192 96 and 48 SNPs were highly concordant with the25k data set (R2 087 in all cases) and all performed betterthan a set of 39 STRs (which had FSTfrac14 018 HO-dogfrac14 05765HO-wolffrac14 04582 supplementary fig S9 SupplementaryMaterial online) commonly used to identify wolf x dog hy-brids in Europe (Randi et al 2014) In particular when fixingan operative individual assignment threshold of qwfrac14 095 forthe identification of hybrids (as currently used in the ongoingEuropean LIFE projects for the management of wolf x doghybrids in Italy) all the top-differentiating SNP sets were ableto correctly recognizegt85 of the hybrids identified with the25 k data set Conversely although based on differentBayesian algorithms assignment results from STRUCTURE on

FIG 6 Principal component analysis (PCA) computed in SVS on the 25 k SNP data set and on the 192 96 48 SNPs with the highest wolf-dog FST

values For each data set PC1 versus PC2 are indicated (axes are not to scale) and the overall FST distribution for the 25 k data set is represented inthe histogram Yellow dots indicate dogs blue dots nonadmixed wolves light blue dots admixed wolves (ADMIXTURE qwlt 0999) The power of thetop 48 SNPs is comparable to that of the 25 k data set indicating that they can be used as reliable ancestry-informative-markers (AIMs) althoughno clear subdivision could be traced between some of the admixed and the nonadmixed wolves

Galaverni et al doi101093molbevmsx169 MBE

2330Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

the 39 STRs identified only 37 of the hybrids detected withthe 25 k SNP data set

GenotypendashPhenotype AssociationsTo identify the genetic bases of three atypical phenotypictraits described in the literature as possible indicators of hy-bridization (Ciucci et al 2003 Anderson et al 2009) we per-formed an exploratory genotypendashphenotype association testWe contrasted the genotypes of nine cases showing blackcoat (BC) phenotypes to 102 wild-type controls The presenceof one or more white claws (WC) was described in 16 caseswhereas 98 controls had all claws with the typical dark greycolor The presence of the spur (SP) on the hind legs wasidentified in five cases compared with 108 controls where thespur was absent The genotypendashphenotype analysis revealeda main peak associated to the black coat color with 90 SNPsabove the significance threshold after Bonferroni correction(fig 7) in an interval of about 12 Mb at the end of chr16 thatincluded the b-defensin gene CBD103 known to be respon-sible for such trait in North American and Italian wolves(Anderson et al 2009 Caniglia et al 2013 Randi et al2014) The alleles of the three most significant SNPs(chr1660391793 chr1661370693 chr1661718721 table 1)were perfectly associated to the phenotypes of all the casesplus another four samples for which the presence of a blackcoat was not known but that carried the KB deletion (supple-mentary table S2 Supplementary Material online) Toevaluate whether the regions containing the SNPs associ-ated with the atypical phenotypes had a wolf or dog an-cestry we proceeded in two ways first we verified their

assignment in PCADMIX for all the cases second we con-catenated all the significantly associated SNPs and recon-structed their haplotype networks The networkreconstructed from the 90 SNPs significantly associatedto the BC showed a single haplotype largely representedin wild-type wolves whereas the six different haplotypesfound in individuals with BC were split into two maingroups both rooted on dog-derived nodes (fig 8)Concordantly the significant SNPs fell into 41 PCADMIX

blocks where all the BC cases had at least one haplotype ofdog ancestry

The genotypendashphenotype association results for the pres-ence of WC revealed a main large peak on chr20 with 74significantly associated SNPs (fig 7) These SNPs spanned aninterval of about 20 Mb which included 286 annotated genessome of which known to be related to pigmentation (sup-plementary file S2 Supplementary Material online) Howevernone of the significant SNPs was associated to all the 16 cases(table 1) but the 22 most significant SNPs on chr20 (betweenchr2027360052 and chr2037651111) clearly explained six ofthem (supplementary table S2 Supplementary Material on-line) Another four cases were associated to a solitary SNP onchr16 (chr1650237955) although with a lower genome-widesignificance (fig 7) The haplotypes reconstructed from thesignificant SNPs on chr20 formed a network centered aroundthe haplotype most commonly observed in wild-type wolves(fig 8) Two haplotypes were associated to the WC pheno-type and formed a distinct branch rooted on a dog nodewhich included two additional admixed individuals for whichno phenotypic information was available In PCADMIX the

FIG 7 Genotypendashphenotype association results for three phenotypic traits considered as potential indicators of hybridization black coat (top)white claws (center) and spur on the hind legs (also known as dewclaw bottom panel) Significance thresholds for the ndashlog10 Chi-Squared P valuesare indicated corresponding to a Bonferroni-corrected probability of 001 (Photo credits from top to bottom Renato Fabbri Willy Reggioni andLuigi Molinari)

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2331Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

FIG 8 Haplotype networks reconstructed from the SNPs significantly associated to the considered phenotypic traits Top-left black coat Top-right white claws Bottom spur (only referred to chr11) Node dimensions are proportional to the number of times each haplotype has beenobserved Haplotypes found in wolves are indicated in blue those found in hybrids (ADMIXTURE qwlt 0999) in light blue those found in dogs inyellow Dashed ovals around nodes identify haplotypes reconstructed from SNPs significantly associated to each phenotypic trait

Table 1 Summary Statistics of the SNPs Most-Significantly Associated with Three Atypical Phenotypic Traits

Trait Top SNPs

Black coat 1660391793(AG) 1661370693(AG) 1661718721(AG)Cases (n frac14 9) 0444 0500 0500Controls (n frac14 102) 0005 0005 0005Chi-squared ndash10 log P 212133 215568 215568Chi-squared Bonferroni P 8531017 3861017 3861017

White claws 2027360052(AG) 2031224886(AT) 2037232759(AG)Cases (n frac14 16) 0188 0188 0188Controls (n frac14 98) 0000 0000 0000Chi-squared ndash10 log P 90931 90931 90931Chi-squared Bonferroni P 000011 000011 000011

Spur 1124498328(AG) 1127397738(AC) 1135263745(AC)Cases (n frac14 5) 0400 0300 0400Controls (n frac14 113) 0000 0000 0000Chi-squared ndash10 log P 201745 152735 201745Chi-squared Bonferroni P 9331016 7431011 9331016

NOTEmdashFor each SNP (indicated as chrposition) the allele associated to the atypical phenotype is showed in italics together with its frequency in ldquocasesrdquo and in ldquocontrolsrdquo andthe association values detected by SVS

Galaverni et al doi101093molbevmsx169 MBE

2332Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

blocks that included the significant SNPs on chr20 showedthat all the six explained cases had at least one dog-derivedhaplotype as for the blocks of the other four cases possiblyexplained by the single significant SNP on chr16

Several regions were associated to the presence of the spurwith a main peak of 145 significant SNPs on chr11 and otherpeaks on a number of other chromosomes mainly chr7chr12 and chr28 (fig 7) In particular the three most signif-icant SNPs lied on chr11 (chr1124498328 chr1135263745chr1141596702 fig 7) and their alleles were perfectly associ-ated to the phenotypes of up to four of the five cases (table 1supplementary table S2 Supplementary Material online) The50-kb regions surrounding the significant SNPs included 467genes (supplementary file S2 Supplementary Material online)enriched for the BP category GO0048856 (anatomical struc-ture development) and a number of HP categories related tophalangeal development syndactyly and polydactyly repre-sented by 11 genes (supplementary file S2 SupplementaryMaterial online) The haplotypes reconstructed from the sig-nificant SNPs in the main peak (chr11) formed a networkwhere the four haplotypes found in the cases were rootedbetween the main wolf node and a wider panel of dogbranches (fig 8) not allowing a plain discrimination of theirorigin However PCADMIX clearly assigned as dog-derived allthe blocks that included the SNPs significantly associated tothe spur

DiscussionAnthropogenic hybridization is considered a global threat tobiodiversity especially in human-dominated contexts wherethe growing diffusion of domestic species might increase therisks of hybridization and introgression in the gene pool ofwild parental populations (Allendorf et al 2001) Ourgenome-wide assessment of hybridization in the Italian wolfpopulation as highlighted by both multivariate and assign-ment procedures showed that the wolf x dog putative hy-brids we analyzed (selected by their low STR assignment oratypical phenotypic traits Randi et al 2014) ranged fromcomplete wolf assignments to c 20 dog-derived genomesOnly a few cases (8 of the hybrids) had a higher dog contentand likely fell within the first three generations from theadmixture

This pattern indicates that recurrent admixture events oc-curred in the past but their legacy has been mostly dilutedthrough time as expected in a selectively neutral perspectivewhere the retained dog ancestry should be lt1 after sevengenerations of backcrossing with wolves

The estimated timing of the hybridization events suggeststhat most of the hybrids do not trace back to the last bot-tleneck of the Italian wolf population but rather to its fol-lowing reexpansion phase In fact even if the first cases weredated in the 1940s and in the 1970s (but not in between) themain peak occurred in the late 1990s when the populationwas expanding in most of its current range (Lucchini et al2002 Fabbri et al 2007) However the frequency of hybridi-zation events seems to decline in the following decade pos-sibly due to the increasing availability of potential wolf mates

and to the stronger social structuring and stability of packswhich can lower the need for dog contributions to reproduc-tion (parallel to what described by Rutledge et al 2010 be-tween eastern wolves and coyotes) Of course given that 170k SNPs offer only a moderately resolved snapshot on thewhole Canis genome we cannot exclude that the legacy ofmore ancient hybridization events could have been left in thewhole population PCADMIX is indeed agnostic on the numberof admixture events that may have occurred as it simplyidentified the ldquodogrdquo blocks within the genomes of hybridsHowever when applying ALDER or similar dating methodsthese models are constrained to pick up either the majoradmixture event (if a main one occurred) the midpoint (incase of continuous admixture events) or the latest event (ifthese were punctuated) The future comparison of multiplecomplete genomes from the Italian and other European pop-ulations will be needed to shed more light on this possibility

From a geographical point of view results show that thefirst hybridization events might have occurred in the mainpopulation refugia in central and southern Italy followed bymore frequent events in the northern Apennines and finallyin human-dominated landscapes along the Tyrrhenian andAdriatic coasts likely surfing the main population expansionwave around the 1990s In particular the Maremma region(western Tuscany) confirmed to be a local hotspot of hybrid-ization as previously described for the wolf (Caniglia et al2013) and other mammals such as the wildcat (Mattucciet al 2013) and the roe deer (Mucci et al 2012) Howevermapping the individual sampling locations cannot properlytake into account the movements and dispersals of individ-uals through the generations following the hybridizationevent

The prevalent dog contribution to admixture appears tocome from German Shepherds which interestingly repre-sented the most common breed in Italy in the last decades(Ente Nazionale della Cinofilia Italiana httpwwwenciit lastaccessed March 1 2017) and is also consistent with the pre-sumably higher probability of successful mating of wolveswith wolf-sized dogs compared with other breeds

Ancestry mapping enabled us to identify several regionswith significant excess of dog or wolf alleles and with steeperor milder genetic clines in the hybrids which are discussedinto details in supplementary text S1 SupplementaryMaterial online

The phenotypendashgenotype association tests on the blackcoat used as a control trait confirmed the ability of our dataset to correctly identify the genomic region hosting the caus-ative mutation for such phenotype although the exclusion ofdogs from the data set due to missing phenotypic informa-tion might have reduced the power of the test Interestinglyboth the local genome ancestry analysis and the haplotypenetworks revealed the dog derivation of this region in all theanalyzed cases confirming previous hypotheses derived fromNorth American wolves (Anderson et al 2009) However theextension of dog-derived haplotype blocks and their rootingwithin the network suggest the occurrence of at least twoseparate events The most ancient one included seven indi-viduals sampled in the Northern Apennines only one

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2333Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

blocks that included the significant SNPs on chr20 showedthat all the six explained cases had at least one dog-derivedhaplotype as for the blocks of the other four cases possiblyexplained by the single significant SNP on chr16

Several regions were associated to the presence of the spurwith a main peak of 145 significant SNPs on chr11 and otherpeaks on a number of other chromosomes mainly chr7chr12 and chr28 (fig 7) In particular the three most signif-icant SNPs lied on chr11 (chr1124498328 chr1135263745chr1141596702 fig 7) and their alleles were perfectly associ-ated to the phenotypes of up to four of the five cases (table 1supplementary table S2 Supplementary Material online) The50-kb regions surrounding the significant SNPs included 467genes (supplementary file S2 Supplementary Material online)enriched for the BP category GO0048856 (anatomical struc-ture development) and a number of HP categories related tophalangeal development syndactyly and polydactyly repre-sented by 11 genes (supplementary file S2 SupplementaryMaterial online) The haplotypes reconstructed from the sig-nificant SNPs in the main peak (chr11) formed a networkwhere the four haplotypes found in the cases were rootedbetween the main wolf node and a wider panel of dogbranches (fig 8) not allowing a plain discrimination of theirorigin However PCADMIX clearly assigned as dog-derived allthe blocks that included the SNPs significantly associated tothe spur

DiscussionAnthropogenic hybridization is considered a global threat tobiodiversity especially in human-dominated contexts wherethe growing diffusion of domestic species might increase therisks of hybridization and introgression in the gene pool ofwild parental populations (Allendorf et al 2001) Ourgenome-wide assessment of hybridization in the Italian wolfpopulation as highlighted by both multivariate and assign-ment procedures showed that the wolf x dog putative hy-brids we analyzed (selected by their low STR assignment oratypical phenotypic traits Randi et al 2014) ranged fromcomplete wolf assignments to c 20 dog-derived genomesOnly a few cases (8 of the hybrids) had a higher dog contentand likely fell within the first three generations from theadmixture

This pattern indicates that recurrent admixture events oc-curred in the past but their legacy has been mostly dilutedthrough time as expected in a selectively neutral perspectivewhere the retained dog ancestry should be lt1 after sevengenerations of backcrossing with wolves

The estimated timing of the hybridization events suggeststhat most of the hybrids do not trace back to the last bot-tleneck of the Italian wolf population but rather to its fol-lowing reexpansion phase In fact even if the first cases weredated in the 1940s and in the 1970s (but not in between) themain peak occurred in the late 1990s when the populationwas expanding in most of its current range (Lucchini et al2002 Fabbri et al 2007) However the frequency of hybridi-zation events seems to decline in the following decade pos-sibly due to the increasing availability of potential wolf mates

and to the stronger social structuring and stability of packswhich can lower the need for dog contributions to reproduc-tion (parallel to what described by Rutledge et al 2010 be-tween eastern wolves and coyotes) Of course given that 170k SNPs offer only a moderately resolved snapshot on thewhole Canis genome we cannot exclude that the legacy ofmore ancient hybridization events could have been left in thewhole population PCADMIX is indeed agnostic on the numberof admixture events that may have occurred as it simplyidentified the ldquodogrdquo blocks within the genomes of hybridsHowever when applying ALDER or similar dating methodsthese models are constrained to pick up either the majoradmixture event (if a main one occurred) the midpoint (incase of continuous admixture events) or the latest event (ifthese were punctuated) The future comparison of multiplecomplete genomes from the Italian and other European pop-ulations will be needed to shed more light on this possibility

From a geographical point of view results show that thefirst hybridization events might have occurred in the mainpopulation refugia in central and southern Italy followed bymore frequent events in the northern Apennines and finallyin human-dominated landscapes along the Tyrrhenian andAdriatic coasts likely surfing the main population expansionwave around the 1990s In particular the Maremma region(western Tuscany) confirmed to be a local hotspot of hybrid-ization as previously described for the wolf (Caniglia et al2013) and other mammals such as the wildcat (Mattucciet al 2013) and the roe deer (Mucci et al 2012) Howevermapping the individual sampling locations cannot properlytake into account the movements and dispersals of individ-uals through the generations following the hybridizationevent

The prevalent dog contribution to admixture appears tocome from German Shepherds which interestingly repre-sented the most common breed in Italy in the last decades(Ente Nazionale della Cinofilia Italiana httpwwwenciit lastaccessed March 1 2017) and is also consistent with the pre-sumably higher probability of successful mating of wolveswith wolf-sized dogs compared with other breeds

Ancestry mapping enabled us to identify several regionswith significant excess of dog or wolf alleles and with steeperor milder genetic clines in the hybrids which are discussedinto details in supplementary text S1 SupplementaryMaterial online

The phenotypendashgenotype association tests on the blackcoat used as a control trait confirmed the ability of our dataset to correctly identify the genomic region hosting the caus-ative mutation for such phenotype although the exclusion ofdogs from the data set due to missing phenotypic informa-tion might have reduced the power of the test Interestinglyboth the local genome ancestry analysis and the haplotypenetworks revealed the dog derivation of this region in all theanalyzed cases confirming previous hypotheses derived fromNorth American wolves (Anderson et al 2009) However theextension of dog-derived haplotype blocks and their rootingwithin the network suggest the occurrence of at least twoseparate events The most ancient one included seven indi-viduals sampled in the Northern Apennines only one

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2333Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

detected as a hybrid by ADMIXTURE and the other six com-pletely assigned to the wolf cluster (qwfrac14 100) indicating agenetic legacy not detectable anymore at the genome-widelevel Conversely the most recent event traced back 3ndash8 gen-erations and the corresponding haplotypes were mostly de-tected in hybrids from the Maremma region(0841lt qwlt 0978) where a high frequency of black-coated canids was recently documented (Caniglia et al 2013)

Also for the white claws we identified a single highly asso-ciated peak on chr20 though relatively large (c 20 Mb) andnot explaining all the cases This region hosted several genesrelated to pigmentation such as MITF (microphthalmia-as-sociated transcription factor) EOGT (EGF domain-specific O-linked N-acetylglucosamine transferase) PRICKLE2 (pricklehomolog 2) WNT5A (wingless-type MMTV integration sitefamily member 5A) and GNAI2 (guanine nucleotide bindingprotein alpha inhibiting activity) However a series of succes-sive highly significant SNPs located between 3692 and3765 Mb was found just 70 kb upstream of WTN5A Thisgene belongs to the WNT family whose members encodefor secreted glycoproteins with signaling functions(Bachmann et al 2005) and was described to be targetedby a miRNA (mir-202) highly expressed in the epidermal cellsof white-colored alpacas (Tian et al 2012) However findingthe causal mutations for white-coloring alleles in the claws ofhybrid wolves needs further investigations and the role ofadjacent genes could also concur to the expression of suchtrait In particular MITF (which is located right at the begin-ning of the peak) is regulated by the same miRNA of WNT5A(Tian et al 2012) and is responsible for the major white spot-ting coloration in Boxer dogs (Karlsson et al 2007) whereasPRICLKE2 is involved in the planar cell polarity signaling path-way which controls the differentiation of follicles and thusaffects the patterning of the skin and of its keratin annexes(Chen and Chuong 2012) Again the reconstructed haplotypenetwork and the local ancestry analysis revealed that in all theexplained cases the region hosting the significantly associatedSNPs was clearly dog-derived confirming previous sugges-tions based on morphological bases (Ciucci et al 2003)Coherently all the cases explained by the associated variantswere detected as possible hybrids (0781lt qwlt 0993) andtheir estimated time of admixture traced back 4ndash9generations

A more complex pattern emerged from the analyses of thegenetic variants associated with the spur showing a numberof peaks on several chromosomes none of them able to ex-plain all the cases per se However this is not surprising since anumber of different genes are known to play a role in thedevelopment of polydactyly both in humans (Biesecker 2011)and in dogs where it occurs more commonly in large breedssuch as Saint Bernard and Bernese (Galis et al 2001) Theregions hosting the significantly associated SNPs includedseveral of such genes In particular PITX1 (paired-like homeo-domain 1) on chr11 and BMP7 (bone morphogenetic protein7) on chr24 are known to be implicated in the developmentof polydactyly in a number of species (Galis et al 2001 Marcil2003 Klopocki et al 2012) similar to LMBR1 (limb develop-ment membrane protein 1) on chr16 that Park et al (2008)

demonstrated to be responsible for such trait in a Korean dogbreed However genes included in other significantly associ-ated regions such as FGF14 (fibroblast growth factor 14) onchr22 and WNT8B (wingless-type MMTV integration sitefamily member 8A) on chr28 could also play a significantrole in the complex development of canine polydactyly(Towle and Breur 2004) Again although the NETWORK resultsare not conclusive from PCADMIX the main associated regionappears to be dog-derived in the analyzed individuals as con-firmed by their low assignment values (0781lt qwlt 0862)and compatible with admixture events occurred from 5 to 9generations in the past

Therefore the three atypical phenotypic traits we focusedon seem to have genetic bases and their presence in wild-living canids is likely a signal of introgression from the dog tothe wolf gene pool although several caveats should be con-sidered First not all the cases of white claws are explained bythe associated genomic variants therefore their presence can-not be considered an absolute signal of hybridization but itcould be due to environmental factors (eg nutrition dis-eases carcass degradation or loss of the external keratin shell)Second several individuals carrying a black coat could not beidentified as hybrids even by genome-wide assignment pro-cedures carried out with thousands of markers This indicatesthat the hybridization event originating such gene flow oc-curred several generations in the past and most of the dog-derived alleles in such individuals got lost resulting in almostpure wolf genomes

Therefore any classification of hybrids based on sole phe-notypic traits for management purposes (eg for hybrid re-moval) would be highly hazardous Rather we think that thelikely dog ancestry of these traits should be used as a pheno-typic marker for a simpler identification of potential hybridsbut a final assessment of their status should be based oncareful genetic investigations To this purpose the reducedpanels of AIMs we identified turned out to be highly concor-dant with the 25 k SNP data set and appeared to performbetter than previously used microsatellites in the identifica-tion of hybrids although such a direct comparison should betreated with caution given the different programs applied toanalyze these two types of markers Therefore a small panel ofhighly informative SNPs could be applied to more extensivemonitoring plans in areas of supposed or documented hy-bridization through microfluidic or quantitative PCR tech-niques (vonHoldt et al 2013) Such approaches would allowthe cost-effective analysis of dozens of samples and markersat a time (eg 48 samples 48 SNPs) even starting fromnoninvasively collected materials such as faeces (Kraus et al2015 Norman and Spong 2015) However any future man-agement practice for the removal of hybrids should be un-dertaken considering the serious possibility to incur in bothtype I (removing nonadmixed wolves erroneously identifiedas hybrids) and type II errors (not removing hybrid individualsfalsely identified as wolves Allendorf et al 2001 Randi et al2014) Therefore these problems should be carefully evalu-ated especially in a population that is undergoing a recentreexpansion but is still threatened by strong poaching andaccidental killings which can weaken the pack structure and

Galaverni et al doi101093molbevmsx169 MBE

2334Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

further promote hybridization (Rutledge et al 2012)Moreover hybrids originated several generations in the pastcan act as good ecological surrogates of nonadmixed wild-living wolves (Caniglia et al 2013 Bassi et al 2017) as docu-mented also for wolf-coyote crosses (Benson et al 2012) andshould not be necessarily removed (VonHoldt et al 2016a2016b Wayne and Shaffer 2016)

Consequently management actions should be primarilyaimed at reducing the high number of free-ranging dogswithin the current wolf distribution and particularly in theircurrent expansion range (the Alps) as indicated by the latestaction plan for the conservation of the wolf in Italy (Boitaniet al in prep) Secondary attention should be focused onrecent hybrids or hybrid packs in order to prevent the diffu-sion of large dog-derived genomic regions and a lower accep-tance by local communities (M Apollonio personalcommunication)

In conclusion the application of multiple genome-wideancestry reconstruction methods allows to clarify the pat-terns and dynamics of admixture even in highly introgressedpopulations map the genetic bases of phenotypical indicatorsof hybridization identify optimal ancestry-informativemarkers and support appropriate management practices

Materials and Methods

Sample Selection DNA Extraction and SNPGenotypingWe genotyped DNA extracted from blood or muscular tissuesamples from 118 wolves 31 village dogs and 72 putative wolfx dog hybrids Wolves were sampled from the whole speciesrsquodistribution in Italy from 1992 to 2015 Putative hybrids werepreviously identified based on genetic evidences (Randi andLucchini 2002 Caniglia et al 2013 Randi et al 2014) andor onatypical phenotypes such as black coat color white claws orspur on the hind legs (Ciucci et al 2003) Wolves and hybridswere sampled from carcasses and individuals live-trapped forscientific purposes or rescuing operations under the ItalianMinistry of the Environmentrsquos permit Dog blood sampleswere collected in local shelters by veterinary personnel alwaysrespecting animal welfare procedures No animal was hurt norsacrificed for the purposes of this study Genomic DNA wasextracted using a DNeasy Tissue Kit (Qiagen Inc HildenGermany) according to the manufacturerrsquos instructions Wequantified double-stranded DNA concentrations by thePicoGreen assay on a Quant-iT fluorometer (TermoFisherScientific Vantaa Finland) and visually evaluated DNA qualityby 1 agarose gel electrophoresis selecting only samples witha minimum of 50 ng DNA (average DNA concentration6063 6 322 ngml) and showing no signs of degradationWe used the CanineHD BeadChip microarray (Illumina IncSan Diego California USA) to genotype the DNA samples atc 170 k SNPs following the Infinium HD Ultra Assay protocoland calling genotypes with GenomeStudio (httpwwwilluminacomdocumentsproductsdatasheetsdatasheet_genomestudio_softwarepdf last accessed March 1 2017) Wethen added 456 dogs from 30 breeds that were genotypedwith the same microarray in the LUPA project and that were

publicly available (Lequarre et al 2011 Vaysse et al 2011)Although the use of the CanineHD BeadChip mostly devel-oped from known dog variants could introduce a limitedascertainment bias against wolves (VonHoldt et al 2011)our haplotype-based analyses although being based on dogrecombination maps are designed to minimize such an effectwhen detecting dog genomic segments that have intro-gressed into the wolf population

Data FilteringAll genotypes were imported into the SNPampVariant Suite801 (SVS Golden Helix Inc Bozeman MT) and werechecked for the possible presence of pairs of individualswith an identity-by-descent score higher than 05 (closely re-lated individuals) Genotypes were filtered to ensure high callrates (gt 95) per sample and per locus (hereafter quality-pruned data set) after discarding SNPs mapping on chromo-somes X and Y We used PLINK 107 (Purcell et al 2007) tofurther filter out loci in linkage disequilibrium (LD) at thresh-old r2frac14 01 (LD-pruned data set) using the ndashdog option inorder to manage the correct number of chromosomes Wethen developed an ad hoc Unix pipeline (available upon re-quest) to integrate and automate the subsequent analysissteps

Assignment and Admixture TestsWe carried out an exploratory principal component analysis(PCA Novembre and Stephens 2008) for the first five com-ponents in SVS to visualize the distribution of samples in thegenetic space using the additive genetic model (Price et al2006) on the quality-pruned data set We then usedADMIXTURE 123 (Alexander et al 2009) on the 25 k LD-pruned data set to reassign each sample to its populationof origin assuming K values from 1 to 40 The most likelynumber of clusters was identified based on the lowest cross-validation error (Alexander et al 2009) and the results wereplotted in R 302 (wwwr-projectorg last accessed March 12017) Introgression fractions assessed with ADMIXTURE (Kfrac14 2)were then used to select the reference dogs reference wolvesand admixed individuals for all the subsequent analyses (seeresults) Individual assignment values of hybrids were verifiedrunning again the software using the ldquosupervised approachrdquowhich allows to fix the reference populations (Alexander andLange 2011) and confirmed by PCADMIX (Brisbin et al 2012see below) which is more appropriate in evaluating the actualadmixture proportions (Falush et al 2016) We used the f3test in THREEPOP (TREEMIX package 112 Pickrell and Pritchard2012) to formally assess the occurrence of admixture in theItalian wolf population using blocks of 500 SNPs each andconsidering Z-score valueslt3 to be indicative of admix-ture (Pickrell and Pritchard 2012)

Estimating the Timing of AdmixtureWe reconstructed chromosomal haplotypes from the quality-pruned data set in SHAPEIT 2837 (Delaneau et al 2012) withstandard parameters and based on dog recombination mapsderived from Mu~noz-Fuentes et al (2015) which were thebest proxy in absence of any wolf-specific map All

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2335Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

coordinates were referred to the canFam2 dog genomenamely the genomic build that was used to originally designthe CanineHD BeadChip We then estimated the averagetiming of the admixture events using ALDER 103 (Loh et al2013) which exploits information derived from the haplotypestructure and the extent of LD decay among neighboring lociThe putative hybrids were grouped into generation cohortsbased on their sampling dates When in a given cohort thenumber of samples was lt4 we pooled adjacent generationsthen the software was run independently on each cohort orpool of cohorts When the admixture was significant (at Pvalueslt 001) we retained the estimated number of genera-tions since the admixture event The average date of admix-ture for each cohort was then calculated as Yfrac14 sY ndash (nG g)where Y is the inferred year of admixture sY is the averagesampling year of the cohort and nG is the number of gener-ations from the admixture event calculated by the softwareassuming a generation time gfrac14 3 years (Skoglund et al 2011)

Moreover we estimated individual admixture times withthe PCA-based admixture deconvolution approach imple-mented in PCADMIX 10 (Brisbin et al 2012) which assignsthe most likely ancestry of haplotype blocks identified in hy-brid individuals to their parental populations PCADMIX wasrun with blocks of 20 consecutive nonoverlapping SNPs Theaverage genome-wide proportion of blocks assigned to eachreference population was calculated for each sample andcompared with the assignment proportions estimated byADMIXTURE The number of generations since the admixturefor each individual was then estimated based on the numberof switches from dog to wolf ancestry blocks (or viceversa)using the formula developed by Johnson et al (2011) mod-ified according to the dog genome length conditional on theproportion of admixture estimated from PCADMIX The num-ber of generations from admixture was converted into yearssince sampling using the same value of 3 years per generationSummary plots across all samples were then compared withthose obtained from ALDER Finally results for single admixedindividuals were visualized in a map to identify possible geo-graphic patterns of admixture through time

Deviations from Admixture NeutralityAccording to the neutral theory the proportions of localancestry in the hybrid samples should stochastically fluctuatearound the average genome-wide proportion of dog and wolfancestral components However selective pressures acting oneach locus or linkage group can result in some genetic regionsbeing introgressed more frequently than others in the ad-mixed genotypes We exploited recently developed Bayesianstatistical approaches (Alexander et al 2009 Brisbin et al2012 Gompert and Buerkle 2012) to identify genomic regionsthat are significantly more or less frequent than expected bychance in the hybrid individuals identified from ADMIXTURE

First haplotype blocks identified by PCADMIX were rankedaccording to their relative proportion of ldquodogrdquo or ldquowolfrdquo as-signment in the hybrids Blocks falling into the top and bot-tom 1 of the genome-wide distribution were labeled asadmixture outliers and adjacent top-ranking blocks werejoined into single segments These regions were then

compared with those identified in the hybrids with theBayesian Genomic Cline analysis in BGC (Gompert andBuerkle 2012) which retrieves SNPs with excess of ancestryin one of the parental populations (the reference wolves andthe reference dogs identified from ADMIXTURE) compared withrandom expectations (a parameter) as well as those withexcess of or resistance to introgression (szlig parameter) Thesoftware was run after filtering for LD and removing all siteswith missing genotypes for 10000 iterations sampled everyfifth one using the ICARrho model based upon dog recom-bination maps derived from Mu~noz-Fuentes et al (2015) inabsence of any wolf-specific recombination map Results weregathered with the estpost utility (Gompert and Buerkle 2012)and the adequate number of burn-in iterations were dis-carded after visual inspection of the log likelihood valuesSNPs were retained as significant when the 95 credible in-tervals of their median a or szlig values did not include zero andtheir median value was included in the top or bottom 1 ofthe genome-wide distribution (Trier et al 2014)

Ancestry-Informative Marker SelectionStarting from the LD-pruned data set we selected three panelsof the 192 96 and 48 most ancestry-informative markers(AIMs) based on their values of FST between dogs and nonad-mixed wolves (with ADMIXTURE qw 0999 see Results) calcu-lated in SVS These three panels of SNPs were used to rerun thePCA and ADMIXTURE analyses and to verify their power to iden-tify wolves dogs and their hybrids Finally their performancewas compared with that of 39 microsatellite markers previouslyused to investigate wolf x dog admixture in Italian andEuropean wolves (Randi et al 2014) based on their STRUCTURE

qw at Kfrac14 2 (run with admixture model independent allelefrequencies 40 k burn-ins and 400 k MCMC for four iterations)

GenotypendashPhenotype Association Testing andIdentification of Trait AncestryAll individuals showing the spur (SP) and white claws (WC)were ranked as ldquocasesrdquo in an exploratory genotypendashpheno-type association study using as ldquocontrolsrdquo all the wild-typewolves and putative hybrids for which phenotypic informa-tion was available Dogs had to be excluded since no pheno-typic information was available for such traits

We performed the association test in SVS (Gorlova et al2011) using the ldquobasic allele modelrdquo in absence of any hy-potheses on the dominance of these traits We then applied aBonferroni correction to identify significantly associated SNPs(at nominal Plt 001) and we plotted the values of the ndashlog10Chi-Squared P for all chromosomes using the GenomeBrowsetool in SVS

To evaluate whether the mutations associated with theatypical phenotypes had wolf or dog ancestry the significantlyassociated SNPs were concatenated and their haplotypeswere reconstructed for each chromosomal region with atleast ten SNPs in PHASE (Stephens and Donnelly 2003) viaDNASP (Librado and Rozas 2009) including the village dogsamples for comparison We then used these haplotypes tobuild median-joining networks in NETWORK 50 (Bandelt et al1999) Finally we used PCADMIX to further evaluate the

Galaverni et al doi101093molbevmsx169 MBE

2336Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

ancestry of these atypical traits by verifying the assignment ofthe haplotype blocks that included the significant SNPs to thewolf or dog clusters

The black coat (BC) phenotype whose genetic basis andlikely dog origin are well known (Candille et al 2007Anderson et al 2009) was used for validation in all theseanalyses with known black individuals used as cases andwild-type ones as controls

Gene Search Gene Ontology Enrichment andFunctional NetworksCoordinates of the outlier blocks from PCADMIX and the out-lier SNPs from BGC as well as the significant SNPs from theassociation tests were converted into the canFam31 assem-bly using the liftover tool (httpsgenomeucsceducgi-binhgLiftOver) in order to exploit the more complete gene an-notation available for such build Information for the genesmapping within 50 kb of distance from the significant SNPs(which is approximately half of the distance between twonearby SNPs of the 25 k data set) was retrieved in EnsemblBiomart (httpwwwensemblorgbiomartmartview lastaccessed March 1 2017) from the Ensembl Genes 84 anno-tation database The genes obtained from each analysis werethen independently checked for possible enrichment towardsspecific Gene Ontology (GO) and Human Phenotypes (HP)categories in G-profiler (Reimand et al 2016) only retainingcategories significant at Plt 005 after BenjaminindashHockbergcorrection Possible functional relationships between theidentified genes were searched in STRING v10 (Szklarczyket al 2015) including all the available types of evidences

Supplementary MaterialSupplementary data are available at Molecular Biology andEvolution online

AcknowledgmentsWe warmly thank all the people who contributed to thecollection of samples in particular D Bigi and M Delogu(University of Bologna) W Reggioni L Molinari and MCanestrini (Wolf Apennine Center) and E Berti (CRASMonte Adone) We specially thank P Milanesi (ISPRA) forproviding the geographical maps of the hybrids L Montana(University of Sherbrooke) for the assistance in blood sampleextraction and D Lawson (University of Bristol) for a prelim-inary exploration of data We also thank M Scandura and MApollonio (University of Sassari) for the useful discussions Weare indebted with the Associate Editor and two anonymousreferees for their constructive and useful comments thatdeeply improved the manuscript

ReferencesAlexander DH Lange K 2011 Enhancements to the ADMIXTURE algorithm

for individual ancestry estimation BMC Bioinformatics 12246Alexander DH Novembre J Lange K 2009 Fast model-based estimation

of ancestry in unrelated individuals Genome Res 191655ndash1664Allendorf FW Leary RF Spruell P Wenburg JK 2001 The problems with

hybrids setting conservation guidelines Trends Ecol Evol 16613ndash622

Anderson TM VonHoldt BM Candille SI Musiani M Greco C StahlerDR Smith DW Padhukasahasram B Randi E Leonard JA et al 2009Molecular and evolutionary history of melanism in North Americangray wolves Science 3231339ndash1343

Arnold ML Martin NH 2010 Hybrid fitness across time and habitatsTrends Ecol Evol 25530ndash536

Bachmann IM Straume O Puntervoll HE Kalvenes MB Akslen LA 2005Importance of P-cadherin b-catenin and Wnt5aFrizzled for pro-gression of melanocytic tumors and prognosis in cutaneous mela-noma Clin Cancer Res 118606ndash8614

Bailey RI Eroukhmanoff F Saeligtre PG 2013 Hybridization and genomeevolution II mechanisms of species divergence and their effects onevolution in hybrids Curr Zool 59675ndash685

Bandelt HJ Forster P Rohl A 1999 Median-joining networks for inferringintraspecific phylogenies Mol Biol Evol 1637ndash48

Bassi E Canu A Firmo I Mattioli L Scandura M Apollonio M 2017Trophic overlap between wolves and free-ranging wolf dog hy-brids in the Apennine Mountains Italy Global Ecol Conserv 939ndash49

Benson JF Patterson BR Wheeldon TJ 2012 Spatial genetic and mor-phologic structure of wolves and coyotes in relation to environmen-tal heterogeneity in a Canis hybrid zone Mol Ecol 215934ndash5954

Biesecker LG 2011 Polydactyly how many disorders and how manygenes 2010 update Dev Dyn 240931ndash942

Brennan AC Woodward G Seehausen O Munoz-Fuentes V MoritzC Guelmami A Abbott RJ Edelaar P 2015 Hybridization due tochanging species distributions adding problems or solutions toconservation of biodiversity during global change Evol Ecol Res16475ndash491

Brisbin A Bryc K Byrnes J Zakharia F Omberg L Degenhardt J ReynoldsA Ostrer H Mezey JG Bustamante CD 2012 PCAdmix principalcomponents-based assignment of ancestry along each chromosomein individuals with admixed ancestry from two or more populationsHum Biol 84343ndash364

Candille SI Kaelin CB Cattanach BM Yu B Thompson D a Nix M aKerns J a Schmutz SM Millhauser GL Barsh GS 2007 A b-defensinmutation causes black coat color in domestic dogs Science3181418ndash1423

Caniglia R Fabbri E Greco C Galaverni M Manghi L Boitani L Sforzi ARandi E 2013 Black coats in an admixed wolf dog pack is mel-anism an indicator of hybridization in wolves Eur J Wildl Res59543ndash555

Caniglia R Fabbri E Greco C Galaverni M Randi E 2010 Forensic DNAagainst wildlife poaching identification of a serial wolf killing in ItalyForensic Sci Int Genet 4334ndash338

Chen J Chuong CM 2012 Patterning skin by planar cell polarity themulti-talented hair designer Exp Dermatol 2181ndash85

Ciucci P Lucchini V Boitani L Randi E 2003 Dewclaws in wolves asevidence of admixed ancestry with dogs Can J Zool 812077ndash2081

Coulson T MacNulty DR Stahler DR vonHoldt B Wayne RK Smith DW2011 Modeling effects of environmental change on wolf populationdynamics trait evolution and life history Science 3341275ndash1278

Currat M Excoffier L 2011 Strong reproductive isolation between hu-mans and Neanderthals inferred from observed patterns of intro-gression Proc Natl Acad Sci U S A 10815129ndash15134

Currat M Ruedi M Petit RJ Excoffier L 2008 The hidden side of inva-sions massive introgression by local genes Evolution 621908ndash1920

Delaneau O Marchini J Zagury J-F 2012 A linear complexity phasingmethod for thousands of genomes Nat Methods 9179ndash181

Deng H Tan T Yuan L 2015 Advances in the molecular genetics of non-syndromic polydactyly Expert Rev Mol Med 17e18

Fabbri E Miquel C Lucchini V Santini A Caniglia R Duchamp C WeberJM Lequette B Marucco F Boitani L et al 2007 From the Apenninesto the Alps colonization genetics of the naturally expanding Italianwolf (Canis lupus) population Mol Ecol 161661ndash1671

Falush D van Dorp L Lawson D 2016 A tutorial on how (not) to over-interpret StructureAdmixture bar plots BioRxiv 066431

Fredrickson RJ Hedrick PW 2006 Dynamics of hybridization and intro-gression in red wolves and coyotes Conserv Biol 201272ndash1283

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2337Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

Freedman AH Gronau I Schweizer RM Ortega-Del Vecchyo D Han ESilva PM Galaverni M Fan Z Marx P Lorente-Galdos B et al 2014Genome sequencing highlights the dynamic early history of dogsPLoS Genet 10e1004016

Fulgione D Rippa D Buglione M Trapanese M Petrelli S Maselli V 2016Unexpected but welcome Artificially selected traits may increasefitness in wild boar Evol Appl 9769ndash776

Galaverni M Caniglia R Fabbri E Milanesi P Randi E 2016 One no oneor one hundred thousand how many wolves are there currently inItaly Mammal Res 6113ndash24

Galis F van Alphen JJM Metz JAJ 2001 Why five fingers Evolutionaryconstraints on digit numbers Trends Ecol Evol 16637ndash646

Galov A Fabbri E Caniglia R Arbanasic H Lapalombella S Florijancic TBoskovic I Galaverni M Randi E 2015 First evidence of hybridizationbetween golden jackal (Canis aureus) and domestic dog (Canisfamiliaris) as revealed by genetic markers R Soc Open Sci 2150450

Gese E Knowlton F Adams J Beck K 2015 Managing hybridization of arecovering endangered species the red wolf Canis rufus as a casestudy Curr Zool 61191ndash205

Gompert Z Buerkle CA 2012 bgc software for Bayesian estimation ofgenomic clines Mol Ecol Resour 121168ndash1176

Gompert Z Buerkle CA 2016 What if anything are hybrids enduringtruths and challenges associated with population structure and geneflow Evol Appl 9909ndash923

Gorlova O Martin J-E Rueda B Koeleman BPC Ying J Teruel M Diaz-Gallo L-M Broen JC Vonk MC Simeon CP et al 2011 Identificationof novel genetic markers associated with clinical phenotypes of sys-temic sclerosis through a genome-wide association strategy PLoSGenet 7e1002178

Grossen C Keller L Biebach I Zhang W Tosser-Klopp G Ajmone PAmills M Boitard S Chen W Cheng S et al 2014 Introgression fromdomestic goat generated variation at the major histocompatibilitycomplex of Alpine ibex PLoS Genet 10d459ndashd467

Hedrick PW Smith DW Stahler DR 2016 Negative-assortative matingfor color in wolves Evolution 70757ndash766

Hovick SM Whitney KD 2014 Hybridisation is associated with increasedfecundity and size in invasive taxa Meta-analytic support for thehybridisation-invasion hypothesis Ecol Lett 171464ndash1477

Imbert C Caniglia R Fabbri E Milanesi P Randi E Serafini M Torretta EMeriggi A 2016 Why do wolves eat livestock Biol Conserv195156ndash168

Johnson NA Coram MA Shriver MD Romieu I Barsh GS London SJTang H 2011 Ancestral components of admixed genomes in aMexican cohort PLoS Genet 7e1002410

Karlsson EK Baranowska I Wade CM Salmon Hillbertz NHC Zody MCAnderson N Biagi TM Patterson N Pielberg GR Kulbokas EJ et al2007 Efficient mapping of mendelian traits in dogs throughgenome-wide association Nat Genet 391321ndash1328

Klopocki E Keuroahler C Foulds N Shah H Joseph B Vogel H Luttgen S BaldR Besoke R Held K et al 2012 Deletions in PITX1 cause a spectrumof lower-limb malformations including mirror-image polydactylyEur J Hum Genet 20705ndash708

Koblmuller S Nord M Wayne RK Leonard JA 2009 Origin and status ofthe Great Lakes wolf Mol Ecol 182313ndash2326

Kraus RHS VonHoldt B Cocchiararo B Harms V Bayerl H Kuhn RForster DW Fickel J Roos C Nowak C 2015 A single-nucleotidepolymorphism-based approach for rapid and cost-effective geneticwolf monitoring in Europe based on noninvasively collected sam-ples Mol Ecol Resour 15295ndash305

Laikre L Schwartz MK Waples RS Ryman N 2010 Compromising ge-netic diversity in the wild Unmonitored large-scale release of plantsand animals Trends Ecol Evol 25520ndash529

Larsen PA Marchan-Rivadeneira MR Baker RJ 2010 Natural hybridiza-tion generates mammalian lineage with species characteristics ProcNatl Acad Sci U S A 10711447ndash11452

Lavrenchenko LA Bulatova NS 2016 The role of hybrid zones in speci-ation a case study on chromosome races of the house mouse Musdomesticus and common shrew Sorex araneus Biol Bull Rev6232ndash244

Lequarre A-S Andersson L Andre C Fredholm M Hitte C Leeb T LohiH Lindblad-Toh K Georges M 2011 LUPA a European initiativetaking advantage of the canine genome architecture for unravellingcomplex disorders in both human and dogs Vet J 189155ndash159

Librado P Rozas J 2009 DnaSP v5 a software for comprehensive analysisof DNA polymorphism data Bioinformatics 251451ndash1452

Lindblad-Toh K Wade CM Mikkelsen TS Karlsson EK Jaffe DB KamalM Clamp M Chang JL Kulbokas EJ Zody MC et al 2005 Genomesequence comparative analysis and haplotype structure of the do-mestic dog Nature 438803ndash819

Loh PR Lipson M Patterson N Moorjani P Pickrell JK Reich D Berger B2013 Inferring admixture histories of human populations using link-age disequilibrium Genetics 1931233ndash1254

Lucchini V Fabbri E Marucco F Ricci S Boitani L Randi E 2002Noninvasive molecular tracking of colonizing wolf (Canis lupus)packs in the western Italian Alps Mol Ecol 11857ndash868

Lucchini V Galov A Randi E 2004 Evidence of genetic distinction andlong-term population decline in wolves (Canis lupus) in the ItalianApennines Mol Ecol 13523ndash536

Mallet J 2008 Hybridization ecological races and the nature of speciesempirical evidence for the ease of speciation Philos Trans R Soc B BiolSci 3632971ndash2986

Marcil A 2003 Pitx1 and Pitx2 are required for development of hindlimbbuds Development 13045ndash55

Mattucci F Oliveira R Bizzarri L Vercillo F Anile S Ragni B Lapini LSforzi A Alves PC Lyons LA et al 2013 Genetic structure of wildcat(Felis silvestris) populations in Italy Ecol Evol 32443ndash2458

Monzon J Kays R Dykhuizen DE 2014 Assessment of coyote-wolf-dogadmixture using ancestry-informative diagnostic SNPs Mol Ecol23182ndash197

Moorjani P Patterson N Hirschhorn JN Keinan A Hao L Atzmon GBurns E Ostrer H Price AL Reich D 2011 The history of African geneflow into Southern Europeans Levantines and Jews PLoS Genet7e1001373

Moura AE Tsingarska E Da browski MJ Czarnomska SD Jedrzejewska BPilot M 2014 Unregulated hunting and genetic recovery from asevere population decline the cautionary case of Bulgarian wolvesConserv Genet 15405ndash417

Mucci N Mattucci F Randi E 2012 Conservation of threatened localgene pools landscape genetics of the Italian roe deer (Capreolus citalicus) populations Evol Ecol Res 14897ndash920

Muhlfeld CC Kovach RP Jones LA Al-Chokhachy R Boyer MC Leary RFLowe WH Luikart G Allendorf FW 2014 Invasive hybridization in athreatened species is accelerated by climate change Nat ClimChange 4620ndash624

Mu~noz-Fuentes V Marcet-Ortega M Alkorta-Aranburu G LindeForsberg C Morrell JM Manzano-Piedras E Soderberg A Daniel KVillalba A Toth A et al 2015 Strong artificial selection in domesticmammals did not result in an increased recombination rate Mol BiolEvol 32510ndash523

Norman AJ Spong G 2015 Single nucleotide polymorphism-based dis-persal estimates using noninvasive sampling Ecol Evol 53056ndash3065

Novembre J Stephens M 2008 Interpreting principal component anal-yses of spatial population genetic variation Nat Genet 40646ndash649

Parham JF Papenfuss TJ Dijk PP van Wilson BS Marte C Schettino LRBrian Simison W 2013 Genetic introgression and hybridization inAntillean freshwater turtles (Trachemys) revealed by coalescentanalyses of mitochondrial and cloned nuclear markers MolPhylogenet Evol 67176ndash187

Park K Kang J Subedi KP Ha J-H Park C 2008 Canine polydactyl mu-tations with heterogeneous origin in the conserved intronic se-quence of LMBR1 Genetics 1792163ndash2172

Patterson N Moorjani P Luo Y Mallick S Rohland N Zhan YGenschoreck T Webster T Reich D 2012 Ancient admixture inhuman history Genetics 1921065ndash1093

Patterson N Petersen DC Ross RE v d Sudoyo H Glashoff RH MarzukiS Reich D Hayes VM 2010 Genetic structure of a unique admixedpopulation implications for medical research Hum Mol Genet19411ndash419

Galaverni et al doi101093molbevmsx169 MBE

2338Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

Pickrell JK Pritchard JK 2012 Inference of population splits and mix-tures from genome-wide allele frequency data PLoS Genet 8e1002967

Pilot M Greco C Vonholdt BM Jedrzejewska B Randi E Jedrzejewski WSidorovich VE Ostrander E a Wayne RK 2014 Genome-wide sig-natures of population bottlenecks and diversifying selection inEuropean wolves Heredity 112428ndash442

Price AL Patterson NJ Plenge RM Weinblatt ME Shadick NA Reich D2006 Principal components analysis corrects for stratification ingenome-wide association studies Nat Genet 38904ndash909

Purcell S Neale B Todd-Brown K Thomas L Ferreira MAR Bender DMaller J Sklar P de Bakker PIW Daly MJ et al 2007 PLINK a tool setfor whole-genome association and population-based linkage analy-ses Am J Hum Genet 81559ndash575

Randi E 2008 Detecting hybridization between wild species and theirdomesticated relatives Mol Ecol 17285ndash293

Randi E Hulva P Fabbri E Galaverni M Galov A Kusak J Bigi D BolfıkovaBC Smetanova M Caniglia R 2014 Multilocus detection of wolf xdog hybridization in italy and guidelines for marker selection PLoSOne 9e86409

Randi E Lucchini V 2002 Detecting rare introgression of domesticdog genes into wild wolf (Canis lupus) populations by Bayesianadmixture analyses of microsatellite variation Conserv Genet 331ndash45

Reimand J Arak T Adler P Kolberg L Reisberg S Peterson H Vilo J 2016gProfilermdasha web server for functional interpretation of gene lists(2016 update) Nucleic Acids Res 44W83ndashW89

Rhymer JM Simberloff D 1996 Extinction By Hybridization andIntrogression Annu Rev Ecol Syst 2783ndash109

Rieseberg LH Whitton J Gardner K Arnold ML Barton NH Hewitt GMBarton NH Hewitt GM Bishop YMM Fienberg SE et al 1999 Hybridzones and the genetic architecture of a barrier to gene flow betweentwo sunflower species Genetics 152367ndash383

Rius M Darling JA 2014 How important is intraspecific genetic admix-ture to the success of colonising populations Trends Ecol Evol29233ndash242

Rosenberg NA Huang L Jewett EM Szpiech ZA Jankovic I Boehnke M2010 Genome-wide association studies in diverse populations NatRev Genet 11356ndash366

Rutledge LY Devillard S Boone JQ Hohenlohe PA White BN Drive EBCanada KJ Biome LD 2015 RAD sequencing and genomic simula-tions resolve hybrid origins within North American Canis Biol Lett1120150303

Rutledge LY Patterson BR Mills KJ Loveless KM Murray DL White BN2010 Protection from harvesting restores the natural social structureof eastern wolf packs Biol Conserv 143332ndash339

Rutledge LY White BN Row JR Patterson BR 2012 Intense harvesting ofeastern wolves facilitated hybridization with coyotes Ecol Evol219ndash33

Saetre G-P 2013 Hybridization is important in evolution but is speci-ation J Evol Biol 26256ndash258

Skoglund P Gotherstrom A Jakobsson M 2011 Estimation of popula-tion divergence times from non-overlapping genomic sequencesexamples from dogs and wolves Mol Biol Evol 281505ndash1517

Stephens M Donnelly P 2003 A comparison of bayesian methods forhaplotype reconstruction from population genotype data Am JHum Genet 731162ndash1169

Stern DL 2013 The genetic causes of convergent evolution Nat RevGenet 14751ndash764

Stronen AV Paquet PC 2013 Perspectives on the conservation of wildhybrids Biol Conserv 167390ndash395

Szklarczyk D Franceschini A Wyder S Forslund K Heller D Huerta-Cepas J Simonovic M Roth A Santos A Tsafou KP et al 2015STRING v10 protein-protein interaction networks integrated overthe tree of life Nucleic Acids Res 43D447ndashD452

Tian X Jiang J Fan R Wang H Meng X He X He J Li H Geng J Yu X et al2012 Identification and characterization of microRNAs in white andbrown alpaca skin BMC Genomics 13555

Todesco M Pascual MA Owens GL Ostevik KL Moyers BT Hubner SHeredia SM Hahn MA Caseys C Bock DG et al 2016 Hybridizationand extinction Evol Appl 9892ndash908

Towle HAM Breur GJ 2004 Dysostoses of the canine and feline appen-dicular skeleton J Am Vet Med Assoc 2251685ndash1692

Trier CN Hermansen JS Saeligtre GP Bailey RI 2014 Evidence forMito-Nuclear and Sex-Linked Reproductive Barriers between theHybrid Italian Sparrow and Its Parent Species PLoS Genet10e1004075

Twyford AD Ennos RA 2012 Next-generation hybridization and intro-gression Heredity 108179ndash189

Veuroaheuroa J-P Primmer CR 2006 Efficiency of model-based Bayesianmethods for detecting hybrid individuals under different hybrid-ization scenarios and with different numbers of loci Mol Ecol1563ndash72

Vaysse A Ratnakumar A Derrien T Axelsson E Rosengren Pielberg GSigurdsson S Fall T Seppeuroaleuroa EH Hansen MST Lawley CT et al 2011Identification of genomic regions associated with phenotypic varia-tion between dog breeds using selection mapping PLoS Genet7e1002316

VonHoldt BM Cahill JA Fan Z Gronau I Robinson J Pollinger JP ShapiroB Wall J Wayne RK 2016a Whole-genome sequence analysis showsthat two endemic species of North American wolf are admixtures ofthe coyote and gray wolf Sci Adv 21ndash13

VonHoldt BM Kays R Pollinger JP Wayne RK 2016b Admixture map-ping identifies introgressed genomic regions in North American ca-nids Mol Ecol 252443ndash2453

VonHoldt BM Pollinger JP Earl DA Knowles JC Boyko AR Parker HGeffen E Pilot M Jedrzejewski W Jedrzejewska B et al 2011 Agenome-wide perspective on the evolutionary history of enigmaticwolf-like canids Genome Res 211294ndash1305

vonHoldt BM Pollinger JP Earl DA Parker HG Ostrander EA WayneRK 2013 Identification of recent hybridization between graywolves and domesticated dogs by SNP genotyping MammGenome 2480ndash88

Vonholdt BM Pollinger JP Lohmueller KE Han E Parker HG Quignon PDegenhardt JD Boyko AR Earl DA Auton A et al 2010 Genome-wide SNP and haplotype analyses reveal a rich history underlyingdog domestication Nature 464898ndash902

Wayne RK Shaffer HB 2016 Hybridization and endangered species pro-tection in the molecular era Mol Ecol 81778ndash793

Zimen E Boitani L 1975 Number and distribution of wolf in ItalyZeitschrift Fur Saugetierkunde 40102ndash112

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2339Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

• msx169-TF1

Pickrell JK Pritchard JK 2012 Inference of population splits and mix-tures from genome-wide allele frequency data PLoS Genet 8e1002967

Pilot M Greco C Vonholdt BM Jedrzejewska B Randi E Jedrzejewski WSidorovich VE Ostrander E a Wayne RK 2014 Genome-wide sig-natures of population bottlenecks and diversifying selection inEuropean wolves Heredity 112428ndash442

Price AL Patterson NJ Plenge RM Weinblatt ME Shadick NA Reich D2006 Principal components analysis corrects for stratification ingenome-wide association studies Nat Genet 38904ndash909

Purcell S Neale B Todd-Brown K Thomas L Ferreira MAR Bender DMaller J Sklar P de Bakker PIW Daly MJ et al 2007 PLINK a tool setfor whole-genome association and population-based linkage analy-ses Am J Hum Genet 81559ndash575

Randi E 2008 Detecting hybridization between wild species and theirdomesticated relatives Mol Ecol 17285ndash293

Randi E Hulva P Fabbri E Galaverni M Galov A Kusak J Bigi D BolfıkovaBC Smetanova M Caniglia R 2014 Multilocus detection of wolf xdog hybridization in italy and guidelines for marker selection PLoSOne 9e86409

Randi E Lucchini V 2002 Detecting rare introgression of domesticdog genes into wild wolf (Canis lupus) populations by Bayesianadmixture analyses of microsatellite variation Conserv Genet 331ndash45

Reimand J Arak T Adler P Kolberg L Reisberg S Peterson H Vilo J 2016gProfilermdasha web server for functional interpretation of gene lists(2016 update) Nucleic Acids Res 44W83ndashW89

Rhymer JM Simberloff D 1996 Extinction By Hybridization andIntrogression Annu Rev Ecol Syst 2783ndash109

Rieseberg LH Whitton J Gardner K Arnold ML Barton NH Hewitt GMBarton NH Hewitt GM Bishop YMM Fienberg SE et al 1999 Hybridzones and the genetic architecture of a barrier to gene flow betweentwo sunflower species Genetics 152367ndash383

Rius M Darling JA 2014 How important is intraspecific genetic admix-ture to the success of colonising populations Trends Ecol Evol29233ndash242

Rosenberg NA Huang L Jewett EM Szpiech ZA Jankovic I Boehnke M2010 Genome-wide association studies in diverse populations NatRev Genet 11356ndash366

Rutledge LY Devillard S Boone JQ Hohenlohe PA White BN Drive EBCanada KJ Biome LD 2015 RAD sequencing and genomic simula-tions resolve hybrid origins within North American Canis Biol Lett1120150303

Rutledge LY Patterson BR Mills KJ Loveless KM Murray DL White BN2010 Protection from harvesting restores the natural social structureof eastern wolf packs Biol Conserv 143332ndash339

Rutledge LY White BN Row JR Patterson BR 2012 Intense harvesting ofeastern wolves facilitated hybridization with coyotes Ecol Evol219ndash33

Saetre G-P 2013 Hybridization is important in evolution but is speci-ation J Evol Biol 26256ndash258

Skoglund P Gotherstrom A Jakobsson M 2011 Estimation of popula-tion divergence times from non-overlapping genomic sequencesexamples from dogs and wolves Mol Biol Evol 281505ndash1517

Stephens M Donnelly P 2003 A comparison of bayesian methods forhaplotype reconstruction from population genotype data Am JHum Genet 731162ndash1169

Stern DL 2013 The genetic causes of convergent evolution Nat RevGenet 14751ndash764

Stronen AV Paquet PC 2013 Perspectives on the conservation of wildhybrids Biol Conserv 167390ndash395

Szklarczyk D Franceschini A Wyder S Forslund K Heller D Huerta-Cepas J Simonovic M Roth A Santos A Tsafou KP et al 2015STRING v10 protein-protein interaction networks integrated overthe tree of life Nucleic Acids Res 43D447ndashD452

Tian X Jiang J Fan R Wang H Meng X He X He J Li H Geng J Yu X et al2012 Identification and characterization of microRNAs in white andbrown alpaca skin BMC Genomics 13555

Todesco M Pascual MA Owens GL Ostevik KL Moyers BT Hubner SHeredia SM Hahn MA Caseys C Bock DG et al 2016 Hybridizationand extinction Evol Appl 9892ndash908

Towle HAM Breur GJ 2004 Dysostoses of the canine and feline appen-dicular skeleton J Am Vet Med Assoc 2251685ndash1692

Trier CN Hermansen JS Saeligtre GP Bailey RI 2014 Evidence forMito-Nuclear and Sex-Linked Reproductive Barriers between theHybrid Italian Sparrow and Its Parent Species PLoS Genet10e1004075

Twyford AD Ennos RA 2012 Next-generation hybridization and intro-gression Heredity 108179ndash189

Veuroaheuroa J-P Primmer CR 2006 Efficiency of model-based Bayesianmethods for detecting hybrid individuals under different hybrid-ization scenarios and with different numbers of loci Mol Ecol1563ndash72

Vaysse A Ratnakumar A Derrien T Axelsson E Rosengren Pielberg GSigurdsson S Fall T Seppeuroaleuroa EH Hansen MST Lawley CT et al 2011Identification of genomic regions associated with phenotypic varia-tion between dog breeds using selection mapping PLoS Genet7e1002316

VonHoldt BM Cahill JA Fan Z Gronau I Robinson J Pollinger JP ShapiroB Wall J Wayne RK 2016a Whole-genome sequence analysis showsthat two endemic species of North American wolf are admixtures ofthe coyote and gray wolf Sci Adv 21ndash13

VonHoldt BM Kays R Pollinger JP Wayne RK 2016b Admixture map-ping identifies introgressed genomic regions in North American ca-nids Mol Ecol 252443ndash2453

VonHoldt BM Pollinger JP Earl DA Knowles JC Boyko AR Parker HGeffen E Pilot M Jedrzejewski W Jedrzejewska B et al 2011 Agenome-wide perspective on the evolutionary history of enigmaticwolf-like canids Genome Res 211294ndash1305

vonHoldt BM Pollinger JP Earl DA Parker HG Ostrander EA WayneRK 2013 Identification of recent hybridization between graywolves and domesticated dogs by SNP genotyping MammGenome 2480ndash88

Vonholdt BM Pollinger JP Lohmueller KE Han E Parker HG Quignon PDegenhardt JD Boyko AR Earl DA Auton A et al 2010 Genome-wide SNP and haplotype analyses reveal a rich history underlyingdog domestication Nature 464898ndash902

Wayne RK Shaffer HB 2016 Hybridization and endangered species pro-tection in the molecular era Mol Ecol 81778ndash793

Zimen E Boitani L 1975 Number and distribution of wolf in ItalyZeitschrift Fur Saugetierkunde 40102ndash112

Genomic and Phenotypic Traces of Admixture in a Hybridizing Wolf Population doi101093molbevmsx169 MBE

2339Downloaded from httpsacademicoupcommbearticle-abstract34923243855352by Aalborg University Library useron 23 April 2018

• msx169-TF1