June 2012 - Weekly Pipeline Set-up

Back to DictyBase

Goal: To set up a weekly CCC pipeline for dictyBase.

• Proposed workflow from Petra
  • On a weekly basis:
    • Monday - dictyBase will identify PDFs for curation
    • Tuesday - Textpresso will process new PDFs, perform CCC search, and send source file output to CCC curation form on tazendra
    • CCC curation performed by dictyBase curators will be stored on tazendra and exported to protein2go database at EBI (UniProt-GOA)
  • We will use four Textpresso categories for the searches
    • dicty gene
    • CCC TAIR
    • CCC assay terms
    • CCC verbs
  • For now, the source file format will be the same as what we used for the dicty searches for the BioCreative task.
  • However, we should consider naming the file by the date the search was performed so curators see that on the curation form.
    • For an example, please see: dicty curation form

• Outstanding issues
  • dicty gene names
    • known issues: 'myosin' vs 'myosin IE - Petra's suggestion: remove 'myosin' from the gene list
    • Greek characters in gene names - can this be addressed at all?
  • Textpresso web sites for dicty
    • currently, two web sites: ccc_for_dicty and Textpresso for dicty
      • we should choose one name and stick with it; I would propose we go with www.textpresso.org/dicty, to keep it simple
      • ccc_for_dicty site is currently running the most updated Textpresso software, but I cannot fine the dicty gene or dicty phenotype categories on this site right now
      • we need to clean up the categories on the site we're going to use - make sure we have the correct dicty categories and remove some of the C. elegans-specific categories (for example, WormBase C. elegans phenotype is the second category listed under Biological Concepts while a dicty curator has to search down the list for any dicty-related categories)
  • CCC curation form
    • We will need a better interface for source file display - perhaps a scrolling menu like the category menu on the Textpresso site.
    • We could sort by year, month, week, etc.
  • Output to protein2go tool
    • What file format? GPAD or GAF?
    • Will need to consult with Tony and Rachael on this.