LEGO April 4, 2016

From GO Wiki
Jump to: navigation, search



Software Updates

  • Status of entity validation after cut and paste?


Implicated in - dealing with vagueness about regulation vs central role in a process

In Barcelona we discussed and provisionally agreed a new qualifier (provisionally called implicated_in) for IMP annotations where it is not clear if a gene product regulates a process or participates in it. Given such situations are quite common in annotation, it presumably needs to be dealt with in LEGO If so, do you have an agreed SOP? Can this be reconciled with the qualifier proposal?

Relations Documentation

  • Relations to describe process-process and activity-process
    • causally_upstream_of_or_within (i.e. affects) - most general use
      • causally_upstream_of
        • regulates (+ and -) - sensu stricto
      • occurent_part_of

After discussion, we agree on this strict interpretation of regulates, excluding parts
Consistent with this, we edited the relevant comment on regulates
 and moved regulates under 'causally upstream of or within'
 We also added a continuant part of relation and under  'causally upstream of or within' and 'part of':

Causally_upstream_of_or_within graph, post edit

 Fully implementing the new part relation would be a lot of churn - especially for end users.
 We provisionally agreed to move to using it, but that this would only be visible in go-plus.owl.
 All other files would have the more general relation.  This is safe because these files are pre-reasoned.
 It may also be possible & desirable to move to using this relation under the hood in LEGO sooner.
  • Using qualifier column in the GAF to add these relations to make more explicit statements about the relation between an entity and a biological process
    • Propose that the relationship specifies that the gene product is the enabler (dph)
      • enables part_of a biological process (or enables a process that is causally_within a biological process)
      • enables a molecular function
      • enables a function that occurs_in a cellular component
      • part_of a cellular component
      • enables regulation of
      • enables a process causally_upstream_of
      • enables a process causally_upstream_of_or_within
Discussion: The above set of relations could potentially all be new qualifiers in GAF files - some or all available for annotation.  
 This would certainly give much more flexibility to annotators, but at a cost in increased complexity.
 Paul T: We should concentrate on getting the OWL representation for LEGO right and not worry too much about supporting expressiveness in legacy systems  
 DOS: I think there is strong demand for at least the general case where am - could support this using a simple qualifier - defined as an expansion to OWL representation
 disadvantage is it would be hard to fold these back up for GAF generation from LEGO.
 General agreement to discuss at GO consortium meeting - perhaps with some preliminary discussion between curators.
  • Dealing with legacy annotations
    • IDA vs IMP ?= part_of vs causally_upstream_of_or_within vs regulates (+ or -) sensu stricto
    • Cellular processes vs organismal level processes - would regulates be different for each of these?

Geneva Meeting

  • Noctua workshop