Andy Jenkinson Bosc Das 2008
-
Upload
bosc2008 -
Category
Technology
-
view
537 -
download
1
Transcript of Andy Jenkinson Bosc Das 2008
BioDAS Andy Jenkinson
BioDAS – Andy Jenkinson 7/20/08 2
BioDAS
BioDAS • an Open Bio project • oversees the DAS protocol
DAS • distributed data integration platform • for visual display
BioDAS – Andy Jenkinson 7/20/08 3
BioDAS
BioDAS – Andy Jenkinson 7/20/08 4
BioDAS
BioDAS – Andy Jenkinson 7/20/08 5
DAS Architecture
Web service protocol • client-server model • client executes commands on the server
Defines 3 constraints • communication method (HTTP) • query model (REST) • data model (XML schema)
BioDAS – Andy Jenkinson 7/20/08 9
DAS Design Principles
Principally for display • should be responsive (fast) • targeted queries • lightweight infrastructure
Use existing technologies • don’t reinvent the wheel
BioDAS – Andy Jenkinson 7/20/08 10
DAS Design Principles
Data remains distributed • “live” data • data providers retain responsibility • good for private/changing data • spreads resources
Easy for data providers to implement • simple protocol • “dumb server, clever client”
BioDAS – Andy Jenkinson 7/20/08 11
DAS Software
Server implementations • Dazzle, ProServer (and others) • pluggable architecture • “standard” plugins provided
Client libraries • Dasobert, Bio::Das::Lite • core communication/parsing only
BioDAS – Andy Jenkinson 7/20/08 12
DAS Extensions
Originally designed for genomics
DAS… extended • additional data types • discovery functionality • feature ontology • advanced stylesheets
1.53E specification
BioDAS – Andy Jenkinson 7/20/08 13
Additional data types
• Protein sequences
• Protein 3D structures
• Alignments
• Molecular interactions
• 3D-EM volume maps
BioDAS – Andy Jenkinson 7/20/08 17
Discovery
Public DAS registry • documents capabilities & coordinate systems • servers can be registries • registry can “mirror” servers
Client
Server
Registry
Server
BioDAS – Andy Jenkinson 7/20/08 18
Ontology integration
DAS enforces no semantics • minimal effort for data providers • limits programmatic analysis
Optional ontology for features • adopted by BioSapiens NoE • incorporates SO; PSI-MOD; ECO • fully backwards compatible
BioDAS – Andy Jenkinson 7/20/08 20
DAS/2
History • bifurcation of DAS project • aim to expand features (e.g. writeback)
Currently • initial grant now expired • lacks features from DAS/1 • minimal adoption, especially outside US
BioDAS – Andy Jenkinson 7/20/08 21
DAS/2
Lessons • backwards compatibility essential • geographical divide a challenge
Options • adopt a single realistic strategy • attempt to merge back • incorporate ideas from DAS/2 into DAS/1
BioDAS – Andy Jenkinson 7/20/08 22
DAS/1 Future
Improve cohesion • better documentation • uniform software implementation
Continue to expand • Authentication • Small molecule DAS • Searches? • Writeback?
BioDAS – Andy Jenkinson 7/20/08 23
DAS adoption
Registered public DAS sources
• 401 servers • 53 institutions • 18 countries
BioDAS – Andy Jenkinson 7/20/08 24
Acknowledgements
OBF
Ensembl
BioSapiens
DAS data providers!
Andreas Prlic Rob Finn Eugene Kulesha Hagen Blankenburg Gabby Reeves Henning Hermjakob Tim Hubbard Ewan Birney
BioDAS – Andy Jenkinson 7/20/08 25
More info
http://biodas.org/ http://www.dasregistry.org/
http://www.biojava.org/wiki/Dazzle http://www.sanger.ac.uk/Software/analysis/proserver/