demonstrating why regulation of process needs to be binned with direct for enrichment of process

[ValWood]

cytokinesis is regulated at lots of points by different signalling pathways (septation initiation signalling, regulation of ring assembly, constriction, etc.....

The network of regulation and non-regulation- nearly all components apart from the physical components of the actual ring itself are mainly signalling pathways, and regulators, which localize to the cytokinetic ring

As an approximation, partitioning "regulation of cytokinesis" from "direct annotation to cytokinesis part of terms" would be equivalent to separating this network into 2 modules (the large oval at the bottom vs. the rest, which ate signalling modules.)

[ValWood]

Ad an example of a list which should enrich for cytokinesis (branched cell phenotype is often associated with cytokinesis)

enriching this list: branched_list.txt

using GO term mapper, give lost of informative enrichments including

using Panther errichment gived nothing really for significance, even though 15 genes involved in cytokinesis, or it's regulation are included:

and cytokinesis is way below threhsold

[ValWood]

this is the intersection between cytokinesis and branched in PomBase: 14 | (branched vegetative cell (FYPO:0000015) [single allele genotypes] AND cytokinesis (GO:0000910))

it includes gene products annotated both to "cytokinesis" (and descendants) and to "regulation of cytokinesis" (and descendants)

Our users would absolutely expect to see both in this list if they retrieved genes "involved in cytokinesis"

[ValWood]

Another reason is that "regulation of sub-process" is always part of a process...although we don't model it this way in GO