Ontology meeting 2012-03-14

From GO Wiki
Jump to: navigation, search

1 hour call

MINUTES: DAVID


TASK FOR TODAY

Try and reach a consensus on "regulation by regulation" by having a group session on Protege.

See http://gocwiki.geneontology.org/index.php/Ontology_meeting_2012-02-15 and minutes.

Becky loaded up Protege and we created a 'x by y' term and created a 'y by x' term. We then defined those terms as ('x' AND 'y') ('y' AND 'x'). When we ran the reasoner the directionality the 'by' construct was lost.

To get around this and to express the directionality, we created a new object_property called 'results_in'. 'results_in' was then chained with 'regulates', 'positively_regulates' and 'negatively_regulates'. New logical definitions were created ('x' AND results_in SOME 'y') ('y' AND results_in SOME 'x'). Reasoning over these terms we found that the correct inferences were made and the logical definitions of the terms were different.

Plan

Notes

Previously we had:

[Term]
id: GO:0006985
name: positive regulation of NF-kappaB transcription factor activity by ER overload response
def: "The conversion of inactive NF-kappaB to the active form, thereby allowing it to activate transcription of target genes, as a result of signaling from the endoplasmic reticulum." [GOC:dph, GOC:mah, GOC:tb, PMID:10390516]
synonym: "EOR-mediated NF-kappaB activation" EXACT []
synonym: "EOR-mediated activation of NF-kappaB " EXACT [GOC:dph, GOC:tb]
synonym: "positive regulation of NF-kappaB transcription factor activity by EOR" EXACT [GOC:dph, GOC:tb]
is_a: GO:0051092  ! positive regulation of NF-kappaB transcription factor activity
relationship: part_of GO:0006983 ! ER overload response

I changed the part_of to is_a, as this is what would be inferred in reasoning

See: http://cvsweb.geneontology.org/cgi-bin/cvsweb.cgi/go/ontology/editors/gene_ontology_write.obo.diff?r1=1.2724;r2=1.2725;f=h

Note there in one missing link that will be filled in by the reasoner:

  • GO:0038097-positive regulation of mast cell activation by Fc-epsilon receptor signaling pathway is_a GO:0033005-positive regulation of mast cell activation

Outstanding questions

These terms present a challenge:

[Term]
id: GO:0043006
name: activation of phospholipase A2 activity by calcium-mediated signaling
def: "A series of molecular signals that leads to the upregulation of calcium-dependent phospholipase A2 activity in response to the signal." [GOC:dph, GOC:jl, GOC:tb]
is_a: GO:0019722  ! calcium-mediated signaling
is_a: GO:0032431  ! activation of phospholipase A2 activity

We would add a definition:

[Term]
id: GO:0043006  ! activation of phospholipase A2 activity by calcium-mediated signaling
intersection_of: GO:0019722  ! calcium-mediated signaling
intersection_of: results_in GO:0032431  ! activation of phospholipase A2 activity

i.e. 'calcium-mediated signaling' and results_in 'activation of phospholipase A2 activity'

but this is insufficient in itself to infer that this is a subclass of "activation of phospholipase A2 activity" - this is because "activation of phospholipase A2 activity" lacks a logical definition (as do all the activation terms).

I propose for a first pass we only tackle those that follow the pattern "{negative,positive} regulation of X by Y" and continue to manually edit "{activation,inhibition} of X by Y". Then we should deal with activation/inhibition other regulation subtypes en masse on some future date.