Ontology meeting 2012-06-06
From GO Public
FOLLOW-UP: ChEBI paper
Chris has the master version and he has not had time to do anything with it yet. They are busy at another meeting til the end of the week, so it is unlikely he will get to it.
FOLLOW-UP: TermGenie paper
Everyone can edit still. Please continue.
FOLLOW-UP: EC numbers in GO
There was a call with a few people from RHEA on May 31st, though we'll need to have a follow-up call. Chris, could you please report on the discussion? The GO curators were given an overview of RHEA and Chris gave an overview of how we can automate EC numbers in GO. One of the key things is that in the RHEA model, they specifically represent unidirectional and bidirectional reactions. General strategy is to let RHEA express these as much as possible and then incorporate that. We need to figure out what to do with reactions that are represented in one resource, but not the other. We have about half the number of reactions as they do. We think that we can map to RHEA automatically using logical defs and biopax defs or we can just import RHEA.
- Progress: to investigate using OPPL as a generic tool that can help us with the data processing here. Added to list of ontology processing tasks
Do we need 'carbohydrate derivative metabolism' and such in GO so that we can group the CHO-phosphate terms under these as they will not go under carbohydrate metabolism anymore?
We think that the best course of action for the exist but not entailed links is for a couple of GO curators to work these out with the ChEBI curators. At the 20th meeting Tanya and David will show the EBI curators how they have been working through the list so that someone at the EBI can meet with the ChEBI curators.
Background: for the time being, we've decided to keep all ChEBI synonyms in GO, including those such as SMILES that may be of little use to humans, but possibly more to computers. Any way we could keep them in the background, rather than fully visible in GO browsers?
From a previous email thread:
"At the moment, there is no simple way of filtering the synonyms. Thus you get the full list of all permutations. With 5 suffixes (i.e. for biosynthesis) and 6 synonyms in ChEBI this will lead to 29 synonyms (one is the main label).
Chris proposed that in the future we could add a pre-processing step (e.g. during the bio-chebi generation). There we could tag/add relevant synonyms. This could also handle synonyms for the alternative ChEBI labels, where we use equivalence axioms (e.g., conjugate base/acid)."
These types of synonyms shouldn't be propagated. Can we get ChEBI to mark these as such and then not propagate them. Chris would prefer to have too few and make the process simpler. Ultimately people should just be able to search ChEBI. Heiko will look into filtering them out. For now we will not propagate ChEBI synonyms. It might be useful to send an e-mail to the GO list to enumerate the types of terms that can be generated by TermGenie. Paola will send the e-mails.
FEEDBACK ON term name changes: "hydrogen" to "proton"
Stemming from this SF request:
Shall I change ALL term names of the form "xxx:hydrogen symporter/antiporter" to "xxx:proton..." instead?
(And add synonyms where missing) ~Paola
Paola will change all of the term names and keep the hydrogen term name as a narrow synonym.
What are we going to do with the outstanding items that are assigned to Karen? David will work on those.