Ontology meeting 2018-04-30

From GO Wiki
Jump to navigation Jump to search

Conference Line

Agenda

GH project link

https://github.com/geneontology/go-ontology/projects/1

Editors Discussion

Discussion of slim maintenance: (Pascale) https://docs.google.com/document/d/1KMASaUdOJSwB1EClRQ69RK77Wqc7AdbWa8EcENidLjM/edit



Minutes

  • On call: Barbara, Chris, David H., Eric, Harold, Jim, Judy, Karen, Kimberly, Laurent-Philippe, Pascale, Paul T., Tanya

Outstanding Pull Requests

  • There are a lot; please check
  • If done, then merge
  • Assign a reviewer if waiting for feedback
    • Reviewer is not typically merging, but they could if everything is okay

Slim Maintenance

  • Pascale looked into who is maintaining what slims
  • Information about slims will ultimately go into the subsets.yaml file
  • Need to cross-check what is included in the header of the ontology file and what is tracked in the subsets.yaml file
  • Should GO archive/obsolete slim subsets that are not maintained?
  • What about the prokaryotic subset?
    • This pre-dated taxon constraints, but should perhaps be retired.
    • We need to reach out and make sure that this is not being used for purposes we haven't thought about.
  • AI: Chris/Pascale will add more entries to the subsets.yaml file
    • The subsets.yaml file will ultimately drive what is displayed on the website
  • What is expected from the person/group who actually maintains the slim?
    • Does the person/group need to edit the ontology?
      • They at least need to submit tickets to go-ontology, if needed
      • They at least need to keep the terms up-to-date
  • Any GO editor will be allowed to update any slim file
  • What are the biological guidelines for maintaining a slim?
    • We need to have documentation for this
  • In OWL the correct terminology is 'subset'
    • Not all subsets are slims; the purposes of a slim vs a subset may be different
    • Should we make this distinction clear by using different values for the type field in the subsets.yaml file?
    • If so, how?

Precomposed Terms vs Creating GO terms on-the-fly

  • SynGO has requested some precomposed terms that are similar to some terms that exist in the ontology
  • There is some concern about how this will scale
  • Groups could instead precompose as needed
    • What would happen to these terms?
    • How will they be used for analysis?
    • Will those terms get any kind of ID?
    • What if other groups wanted to use these terms for analysis?
    • What implications does this have for reproducibility?
  • Experts often want very specific terms
    • Some of these have been added to the ontology, but now some of these are being removed
    • What's the best thing to do?
  • AI: Barbara will try to get some feedback from users on terms and enrichment expected from GO users at the UCL workshop

Enrichment Results

  • David H noticing some changes in enrichment analyses
  • Some new enrichment results are showing enrichment to high-level terms that we might not expect to see in enrichment results
    • Is this due to changes in annotation or changes in the ontology?
  • AI: Refactor the cellular process branch of the ontology.

Signaling Pathways

  • When should we create distinct terms for different signaling pathways?
  • When there are different downstream effectors?
  • Specifically wrt regulation terms for the ERBB4 signaling pathway, we will request that the curator annotate to the less specific parent terms
  • AI: Pascale will get feedback on the different ERBB signaling pathway terms and if they are really needed in GO

Assertion that appear and disappear

  • AI: Harold has been put on the case; Val should supply more information about ontology versions
  • It's not clear from the ontology file what is going on
  • The same terms seem to be disappearing

SAB

  • David will present the ontology slides
  • Will not go into great detail on the specific ontology actions, more about priorities and procedures