Annotation Conf. Call 2016-03-08

From GO Wiki
Jump to: navigation, search

Bluejeans URL

Agenda

Ontology Request Guidelines

IC Annotations from ISS Annotations

  • Proposal from Rebecca, UCL

Media:IC_from_ISS_Annotation_Call.docx

Next Annotation Consistency Exercise - March 22, 2016

Minutes

On call:

  • Berkeley: Chris, Moni
  • EBI: Aleks, David OS, Melanie, Paola
  • FlyBase: Giulia, Helen
  • MGI: David H, Li
  • PomBase: Midori
  • SGD: Edith, Stacia
  • UCL: Barbara, Paul, Rachael, Rebecca, Ruth
  • WB: Kimberly
  • Zfin: Sabrina

Ontology Requests

  • There are new guidelines for ontology term requests
  • The guidelines are linked from the Github ontology tracker, specifically from the top of the page when a New Issue is created
  • These guidelines are in place to help the ontology editors so they don't have to re-learn the biology a curator has already investigated to make the term request
  • Please include as much background and examples as you can and where possible, especially with edit requests, include screenshots from AmiGO and/or QuickGO, so that issues can be dealt with efficiently
  • Note that sometimes what are thought to be ontology issues may actually be issues with browser settings
  • There is now a github tag for 'insufficient info' to flag tickets that don't have enough information
  • Please also respond to comments in a timely manner as the ontology editors are having to look at a lot of different topics and it's easier to work on an issue when the subject matter is still fresh in their minds
  • Jane also created Help documentation on the TermGenie site that walks curators through a decision tree wrt when to use TermGenie vs when to create a ticket

IC Annotations from ISS Annotations

  • Background: Sometimes curation seems incomplete when curators are not able to make biologically relevant IC annotations from existing ISS annotations
  • For example, there may be an ISS annotation to regulation of transcription with a Wnt gene in the AE column; this could lead to an IC annotation of regulation of Wnt signaling pathway
  • This annotation issue has been discussed at various meetings (Oregon, 2009; UCL, 2011; Stanford, 2012?? - can't currently find anything in the Stanford minutes) with the practice first being allowed and then later disallowed
  • Being able to make these annotations would help fill in annotation gaps
  • After discussion, it was agreed (and no one objected) to allow the creation of IC annotations based on one or more supporting ISS annotations
  • The issue of what reference to use was then discussed - use an existing GO_REF like GO_REF:0000036 making IC annotations based on two or more existing annotations or something else?
  • We decided that, for clarity, IC annotations from ISS will need a separate reference (not GO_REF:0000036) in case we want to distinguish them at any point
  • David Hill brought up the problem of tracing statements where both the GO term and the annotation extension are used to make the IC annotation (in the first example in the doc the Lmx1b transcription factor is upregulating transcription of the Wnt1 gene)
  • This issue may be true for all IC annotations, not just those derived from ISS annotations
  • Melanie responded that they are working towards a new GPAD format that will allow the source annotation to be traced precisely
  • Ideally we need something like this or to somehow be able to record an annotation ID (not just a GO term) for full traceability
  • But for the short-term curators could add an annotation note in their tool to specify the exact source of the supporting IC evidence
  • Action Item: Rebecca will send a proposed wording for a new ICC-from-ISS GO_REF to the annotation managers, and proposed new IC documentation to add to the website (DONE)

Annotation Extension (AE) Relations

  • Curators should continue to update annotations that use annotation extension relations slated for obsoletion
  • Note that documentation for some AE relations is still incomplete and has led to misuse, e.g. during vs happens_during
  • Action Item: David OS will officially deprecate AE relations that were slated for obsoletion. Annotations that use these relations should then get flagged in the Jenkins QC reports so curators can see which annotations need to be fixed.
  • Action Item: Ruth will check which AE relations are next up for curators to correct

Annotation Consistency Exercise

  • Next paper up is on regulation of C. elegans UNC-104 kinesin motor activity
  • Action Item: Kimberly will send around an annotation spreadsheet for people to fill in their annotations and comments