JBrowse and Inter-"Mine" Communication - IMDEV 2017

JBrowse and Inter-"Mine" Communication - IMDEV 2017

Lightning Talk about InterMine/JBrowse integration and extensions to Inter-"Mine" Communication, presented at the 2017 InterMine Developer Workshop and Hackathon (IMDEV 2017) held at the Joint Genome Institute (JGI) in Walnut Creek, CA

Presented on Thursday, March 30th 2017 by Vivek Krishnakumar

655ece370aa88ec83d11254234ded6ce?s=128

Vivek Krishnakumar

March 30, 2017
Tweet

Transcript

  1. 1.

    JBrowse & Inter-mine communication Vivek Krishnakumar J. Craig Venter Institute

    vkrishna@jcvi.org 2017 InterMine Workshop and Hackathon Thursday, March 30th @ 11:10am
  2. 3.

    InterMine <-> JBrowse integration ➔JBrowse & REST API spec ➔Components

    in InterMine <-> JBrowse integration ➔Issues with initial implementation ➔Fixes implemented in InterMine core and JBrowse
  3. 4.

    JBrowse & REST API spec • JBrowse (http://jbrowse.org) • Javascript

    based portable genome browser tool developed by GMOD community • Ships with two REST Store adapters: Features and Names • REST Features Store handles genomic features within specified region: • :base/stats/global • :base/stats/region/:refseq?start=:start&end=:end • :base/stats/regionFeatureDensities/:refseq?start=:start&end=:end&basesPerBin=20000 • :base/features/:refseq?start=:start&end=:end • REST Names Store handles search and retrieval of features by identifier: • :base/names/?startswith=:searchterm • :base/names/?equals=:searchterm • Accepted response type: application/json http://gmod.org/wiki/JBrowse_Configuration_Guide#Writing_JBrowse-compatible_Web_Services
  4. 5.

    Components in InterMine <-> JBrowse integration • populate-child-features postprocessor task

    • Processes SequenceFeature entities to ensure that feature hierarchy is correctly structured (based on SO) • Genomic feature Engine (API) • Returns genomic feature data in JBrowse compatible JSON format • Config generator service (API) • Returns feature tracks (trackList.json) & reference sequence listing (refSeqs.json) • Names service (API) • Returns list of genomic features matching the input search term • Embedded ReportDisplayer • Injects (via iframe) JBrowse visualization into InterMine report page http://intermine.readthedocs.io/en/latest/webapp/third-party-tools/jbrowse/
  5. 6.

    Issues with initial implementation • populate-child-features postprocessor • Failed to

    correctly construct Gene model JSON object |- gene:1 |----- mRNA:1 |----- exon:1 |----- exon:2 |----- mRNA:2 |----- exon:3 • Feature Engine & Config generator • Failed to properly convert InterMine Class names to SO terms e.g. AntisenseLncRNA → antisense_lncRNA` • Failed to specify transcript type in track configuration • Data issues • Failed to represent Gene model CDS entity as disjointed parts due to CDS loading mechanism e.g. FlyBaseCDSFastaLoader, AraportCDSFastaLoader, ChadoSequenceProcessor • populate-child-features postprocessor • Improved logic populates the childFeatures collection appropriately |- gene:1 |----- mRNA:1 |----- exon:1 |----- exon:2 |----- mRNA:2 |----- exon:3 • Feature Engine & Config generator • Automatically resolves IM Class to SO type • Ability to specify which tracks to display and define extra parameters via web.properties config • Fixes to ProcessedTranscript glyph (in JBrowse) • Automatically infers disjointed CDS parts from exon parts and total CDS span Fixes made to InterMine & JBrowse core intermine/#1454 intermine/#1426 jbrowse/#872 @sergiocontrino @justinccdev @vivekkrish
  6. 7.

    Extending FriendlyMine Functionality ➔Current state of Inter-mine communication based on

    standard Homology data model ➔FriendlyMines extended to support alternate Homology data model(s) ➔Extending FriendlyMines to support new Entity types (e.g. SyntenyBlocks)
  7. 8.

    Inter-mine communication using Homology data 1 2 3 Homologue data

    model Friendly mine config in web.properties
  8. 9.

    Extending FriendlyMines to support alternate Homology model(s) Standard Model Class

    Name Homologue Primary Entity gene Related Entity homologue Relationship type (orthologue, paralogue, LDO, etc.) PhytoMine Data Model Homolog gene1 gene2 type (NO VALUE) Method NA method (one-to-one, one-to- many, many-to-many)
  9. 10.

    FriendlyMines extended to support alternate Homology model(s) 1 2 3

    stable/e93905e @vivekkrish Homolog data model Friendly mine config in web.properties
  10. 14.

    Community Orthologs (Phytozome, LIS) eFP Browser (BAR) Real-time Data Centers

    Community RNA-Seq expression (NCBI SRA) Publications (NCBI) Pathways (KEGG) Mt4.0 • Genome sequence • Genome annotation MedicMine Gene List Analysis Gene Report Query, Web Services MedicMine - Data Warehouse Warehouse Data Sources Mt Gene Indices (DFCI) Krishnakumar et al. 2014, Plant Cell Physiol
  11. 16.

    Community Orthologs (PhytoMine) Co-expression (ATTED) Real-time Data Centers Community Protein

    Interactions (InAct, BioGrid) Gene expression (Araport, BAR) Publications (NCBI, UniProt) Pathways (KEGG) Araport11 ThaleMine Gene List Analysis Gene Report Query, Data Tables Web Services ThaleMine - Data Warehouse Warehouse Araport Data Sources Krishnakumar et al. 2016, Plant Cell Physiol Reannotation of Col-0 genome (Araport)
  12. 17.

    Additions to ThaleMine Heatmap for 113 RNA-Seq Studies Gene Report

    Gene List Analysis Gene Report (e.g. dicer-like 1/AT1G01040) Medicago JGI PhytoMine Human RNA-seq study RNA-seq study Gene Genes
  13. 18.

    Summary • Issues plaguing the functionality of InterMine’s zero-config JBrowse

    integration have been resolved • DISCUSSION: Fold-in changes made to JBrowse code base into InterMine (for maintainability) • FriendlyMine functionality extended by implementing link classes to support alternative Homology data model(s) • FriendlyMine functionality extended further to support Synteny based interlinking of InterMine instances • Araport and MedicagoGenome.org implement InterMine warehouses, ThaleMine and MedicMine, respectively.
  14. 19.

    Acknowledgments • JCVI • Christopher D. Town (PI: Araport, LegFed)

    • Agnes P. Chan (co-PI: Araport, LegFed) • Cambridge • Sergio Contrino (Araport Software Dev) • Gos Micklem (co-PI: Araport) • Justin Clark-Casey • Julie Sullivan • NCGR • Samuel Hokin (LIS Lead Software Dev) • Andrew Farmer (co-PI: LegFed) • JGI Phytozome • Joe Carlson (Lead Software Dev) • InterMine and GMOD community