Ontology meeting 2017-07-07

From GO Wiki
Jump to: navigation, search

Conference Line


This week we will have Pascale go over the work she is doing on targeting.

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


  • On call: Barbara, Chris, David H, David OS, Harold, Jim, Karen, Kimberly, Pascale

Protein Targeting vs Localization

  • Pascale presented slides on the work on these terms
  • Targeting refers to using a molecular signal to effect transport to a specific location
  • Currently, targeting is a child of transport
    • But is this in agreement with the original meaning of these terms?
  • Some annotations to targeting have been made to the protein that is being targeted (i.e. the object of the transport)
  • Precise meaning and correct use of targeting, importing, localizing terms seem to be unclear to curators
  • Github ticket: merged protein targeting to nucleus and protein localization to nucleus
  • Localization seems the broadest term that would include establishment (targeting, transport, import) and maintenance (anchoring)?
    • Need to review term definitions for this discussion
    • Localization defines the endpoint that experimenters look at
    • As a high-level term, localization can be used to annotate when the mechanism is not known, but there is an (phenotypic) effect
    • But we need to be clear, then, on what processes the establishment and maintenance terms refer to
  • Review design patterns for these terms
  • What is the meaning of 'directed movement'? How many terms does that really apply to? If there is only one mechanism, then is transport sufficient and establishment not needed?
  • For targeting terms, we should look at the referenced papers to see if they elucidate the mechanism. From there, decide if the terms should be merged into transport terms?
  • Add annotation comments to illustrate what types of activities/gene products should be annotated to the specific process terms
  • Localization vs location - which makes most sense? Localization is probably the majority usage in the literature, so will stick with that.

Failed Travis Checks

  • How to re-run
    • Follow link to Travis
    • Click on Restart build
  • Note that whether or not this will fix the problem depends on why the build failed

OBO Edit

  • edit and release OBO files don't open in OBO-Edit
  • Is this just the result of a simple header issue? Can it be fixed reasonably quickly?
  • Do we want to continue doing any work in OBO-Edit?
  • AI: David OS will investigate further and then we'll decide what OBO-Edit can/should be used for

Term Obsoletion

  • If you're going to obsolete a term, you need to make sure you've checked where else the term is used
    • Annotations
    • Logical definitions
    • External ontologies?
  • AI: Update documentation to reflect that checking how the term is used should be the first step.
  • Always need to send out an email giving two-weeks notice for term obsoletion.

Documentation Updates

  • We will review in August how to edit the md documentation files in github.
  • Proposal - anyone can update the docs, but one person will be tasked with reviewing the pull requests and approving.

Adding Terms to the Imports

Suggestions for Review

  • Not working? Can't suggest reviewers?

Cell Type-Specific Terms

Issues with Merges

  • Travis build is failing, because the merged term ID appears to still exist in the ontology
  • Need to double-check the order of the merge
  • The losing term had an alt_id of its own
  • Note: don't search Protege with term ids; use an underscore or the IRI fragment
  • Will review this merge together on next Friday's call

August Meeting Agenda