Ontology meeting 2012-01-11
In preparation for the Protege workshop in Hinxton
- Do we need to have a call with all participants before then? We already have a call planned for Thu. Jan. 19th to discuss Paul T's proposal. Otherwise, we can all refer to the Protege workshop wiki for the list of tools we need to have up and running on our laptops.
- We agreed that the GO Editors could take turns in acting as gatekeepers for the new Termgenie, to check terms submitted by curators and by other editors.
Here's a proposal for a rota, that tries to take into account SF triage and GO-help rotas as well as traveling to the Hinxton workshop. Feel free to edit if you're not happy with the schedule:
Week beginning Jan. 9th: Becky
Week beginning Jan. 16th: David
Week beginning Jan. 23rd: Paola
Week beginning Jan. 30th: Jane
Week beginning Feb. 6th: Tanya
- FYI, we've submitted a request in the TG tracker for a template to create chemical-binding terms.
- New TG template for regulation of MF & BP:
A minor point: synonyms for "positive regulation" include "up regulation", but not "up-regulation" nor "upregulation". Same for negative regulation, where the new TG provides only "down regulation" but not "down-regulation" nor "downregulation". This should be fine, because if a user searched GO with "upregulation/up-regulation of ..." he/she should still be able to find the relevant term, and there would be less clutter in the ontology, but just to double-check.
Discussion notes I: regulation_of by regulation_of terms
- It seems that we still need a long-term solution for terms such as the following:
This also relates to many SF requests we have, e.g. "negative regulation of cellular hyperosmotic salinity response by negative regulation of transcription from RNA Polymerase II promoter". These unresolved requests may be retrieved by the following link (which also includes "in response to" terms that are in the agenda for the following call):
We also have this type of request:
We need to agree on a strategy to deal with these. Then, we'd need the relevant templates in the new TG please.
Proposed Solution (Chris)
It's imperative that we capture the essential asymmetry of these terms. An intersection of two regulation terms doesn't work (see the recent problem re phos/kinase activity)
Solution 1 : introduce a sub-relation of regulates to mark out one partner
Solution 2 : make the definition of the form
"regulation of X via Y" EquivalentTo Y and 'has part' some "regulation of X"
(same for neg/pos)
If we also include axioms
has_part o regulates -> regulates has_part o negatively_regulates -> negatively_regulates has_part o positively_regulates -> positively_regulates
Then we end up with the desired(?) inferences, i.e.
"regulation of X via Y" SubClassOf "regulation of X" "regulation of X via Y" SubClassOf Y
This requires "regulation of X" to be pre-composed (TG doesn't yet handle nested expression but it will in the future)
Discussion notes II: "involved_in" versus "during"
- Stemming from this SF discussion:
We also have this request:
- Chris Mungall
- David Hill
- Tanya Berardini
- Jane Lomax
- Rebecca Foulger
- Karen Christie
- Harold Drabkin (part)
TG1 Gatekeeper Rota
- Agreed to go with Paola's rota.
- Gatekeeper should check for pending requests AT LEAST once a week.
- AI: Paola to put rota somewhere visible.
- Done - it's linked in the Ontology Development section of the wiki, under "Other Ontology Development Group Activities": http://gocwiki.geneontology.org/index.php/TermGenie1_Gatekeeper_Rota
- AI: Paola and Becky to send Heiko a list of synonyms for regulation terms in TG (including whether they are BROAD, EXACT etc).
regulation of x by regulation of y
(our favourite topic)
- E.g. regulation of glucose metabolism by regulation of transcription [of gene products that take part in glucose metabolism pathway]
- This overlaps with Paul T's suggestion of 'regulating the AMOUNT of something that is available for function'.
- We discussed Chris's option of using 'regulation of glucose metabolism VIA regulation of transcription'. So far we've been using two is_a parents (e.g. BY) so this is a different way of looking at it. We're not sure it's the correct way of representing it: if the transcription is the WHOLE thing that is regulating glucose metabolism, does it make sense to have this as a HAS_PART relationship. It may be best to talk about this in person at the Protege workshop at the end of Jan 2012.
- AI: David will send round examples from Dianna's SF items.
Preparation for Protege Workshop
- David has let everyone know that if you're attending the Protege workshop, and a relevant discussion item is on the agenda for the Ontology Editors call, then you should all attend the call.
- As it happens, we didn't get time to discuss this, so it's a priority for next weeks meeting (Jan 18th).
During vs Involved In
- We didn't get time to discuss this. Moved to the agenda for next week (Jan 18th). David has some views on this but will hold off on edits until then.