GO-CAM Working Group Call 2018-09-18

From GO Wiki
Jump to: navigation, search

Meeting URL

https://stanford.zoom.us/j/976175422

Agenda

Evidence Codes in Noctua

  • Currently, Noctua allows for use of the full Evidence and Conclusion Ontology
  • GO has typically only used a subset of ECO codes, e.g. IDA
    • ECO:0000314 direct assay evidence used in manual assertion
    • has_related_synonym 'IDA'
    • database cross reference GOECO:IDA (inferred from direct assay)
  • Some groups, e.g. SynGO use an expanded set
  • Have had feedback that it's harder to find the traditional GO codes, although searching by the first word of the label works well for finding the relevant code
  • Need to gauge how big an issue this is for curators
  • There are possible work-arounds but they all have pros/cons
    • Hard-coded list
    • Boost to specific fields in search result, i.e. fine-tuning specific autocompletes - any unintended consequences?
  • Best long-term solution?

Relations between MF and Input(s)

  • has_input vs has_direct_input
  • Proposal: replace has_direct_input with has_input; obsolete has_direct_input
  • Need to review has_input annotations to remove any extensions that are inconsistent with GO-CAM usage, i.e. an indirect or unknown proximity for an input
  • Seth retrieved, as of 2018-07-31, all MF annotations that use has_input in annotation extensions.
    • Initial review:
      • used to capture a regulatory effect, e.g. protein kinase activator activity, when it was not known whether the effect was direct or indirect (e.g. expression of protein or complex X increases the activity of Y)
      • used to capture a regulatory subunit whose presence is necessary for the activity to occur (e.g. cyclin-dependent protein kinase)
      • used to capture an enzymatic activity when it was not known if the effect on a substrate was direct or indirect (e.g. caspase-dependent but not known if it was the caspase mutated)
      • used to capture an enzymatic substrate where there wasn't also a direct binding assay in the paper (e.g. testing possible chemical substrates for glucuronysyltransferase activity)
      • used to capture metal ion-dependence of protein binding (e.g. Ca2+-dependent protein binding)
      • used (correctly) to capture the physiologically relevant input in a binding reaction (i.e. cross-species experiment where with/from captures experimental binding partner and AE the relevant binding partner)
  • Relations Ontology working group (broader than just GO) that is also considering how to model participants in an MF and documentation of has_input and child relations

Minutes

  • On call: Chris G, David, Dustin, Edith, Giulia, Harold, Helen, Jim, Karen, Kevin, Kimberly, Laurent-Philippe, Marie-Claire, Paul T, Rob, Sabrina, Seth, Shur-Jen, Stacia, Suzi A

Evidence Codes in Noctua

  • Decision: leave evidence codes as is in Noctua
  • Will we continue to use full ECO
  • Autocomplete searches perform best when searching on first word in term label, e.g. 'direct' or 'mutant'