Ontology meeting 2013-12-05

From GO Wiki
Jump to: navigation, search

Taxon constraints file

Let's discuss the transition of taxon constraints into editors file. Do we want to carry on generating a separate taxon constraints purl for legacy applications?

Moving taxon constraints to main file is dependent on moving to 'megafile' strategy. 
But we should be careful to make sure that this move doesn't mean dumping all relationships
to regular OWL release file without any announcement.

Jane: do we need to make sure we still make generate a separate taxon constraints file - resolvable by PURL?

Chris: Issues with choice of import file to use.  
Right now we use NCBI taxon slim, but we could choose to generate an NCBI taxon 'slim' based on terms used 

David: What does this mean for separate file for Rachel (as mentioned in her paper)?

Agreed that she should have a purl that, once we've moved to maintaing these in the main file, 
will resolve to README describing how to move to using the main file.

Lag resolving class purls

EFO recently added some terms to GO and they need them to resolve before they can add them to their release. The date OntoBee has for GO is 15:10:2013 - this is probably too much of a lag. Can we ask them to update more often? Or resolve our purls elsewhere?

TODO Make ticket on ontobee asking for GO to be updated more frequently on the site

Protein complex involved in bp terms?

Stemming from this ticket of Birgit's [1]

She's basically asking whether she should add 'x mf complex involved_in bp' as child terms to 'x mf complex' terms, or whether synonyms for the specific subtypes are sufficient on the 'x mf complex'.

There's also an additional question from Birgit about whether to create a class to group complexes where there is no obvious function. The example is a complex involved in the innate immune response. It's known to be nucleotide binding, but that's not really sufficient for a grouping class. Use the process? Or no grouping?

David OS: In vast majority of cases I doubt it is worth making named classes defined as EquivalentTo protein complex that capable_of_part_of some <SubClassOf biological_process>, but it is worth using regular relationships to link complexes to biological processes, just as specific complexes (with defined components) should only be linked to MF via relationships. Good example is STING complex:

Note - planning to discuss this in meeting next week with 
Could we use AmiGO2 to provide a way to query for protein complex via process/function?
Chris: Should be possible.

Chris: We need to be careful about the semantics here.  Putting it in GO means we are saying the complex has this function in all species.  There is also a plan to annotate complexes directly.  The semantics of this annotation will limit assertions to a single species.  
GO eds should probably attend upcoming annotation ID conf call organised by Kimberly VA:

Monday, December 9th, at 5pm GMT/12pm EST/9am PST.

 agenda here:

Stemming from question of adding 'luteinizing hormone' in GO

See email thread. Chris says: "GO eds - we will want to switch IDs in the production logical definitions before this, to check the inferences. That's another external ontology imports file - not a big deal. Discuss on thursday?"

Action Item: Move to using pro for protein, rather than CHEBI.
Do this at same time as moving axioms referencing pro from the extensions file to main file.

Review of template requests in Jira

 Moved to next week - should be sole - or at least first agenda item 

New TG template for 'cell migration'?

 Postponed to next meeting when Paola is around 

Note that 'cell chemotaxis' is_a 'cell migration', which is fine based on their definitions. However, both 'cell chemotaxis' and 'cell migration' terms have logical defs based on alters_location_of, e.g.


id: GO:0072676 ! lymphocyte migration

intersection_of: GO:0016477 ! cell migration

intersection_of: alters_location_of CL:0000542 ! lymphocyte

Chris says: "It would be good to use a more generic relation, but this depends on the hierarchy being disjoint with other classes using the same relation with a different genus.."

Review proposed fixes to terms for extracellular & extra-organismal region terms - including status of 'other organism'

See: https://www.ebi.ac.uk/panda/jira/browse/GO-216

{ DOS: don't detailed notes on this.  Anyone want to add anything? }

General support for proposal except that we need to tweak definitions so that part_of some \<anatomical entity SubClass\> is valid for annotation extensions

TODO: David to write up fix more completely and implement in a branch.

 Only got as far as this, all other items should be moved to future meetings 

Patterns for defining maintenance of location.

Should maintenance of location be a subclass of localization rather than a part? This appears to be consistent with the current textual definition of localiation and would also allow complete automation of new term classification in this branch. Details here: https://www.ebi.ac.uk/panda/jira/browse/GO-232

protein-DNA complexes revisited

We decided some time ago that this was not a useful grouping because it's unclear what constitutes a protein-DNA complex (any protein complex that binds DNA?) making for very tricky manual classification under this term, and we didn't feel it was a useful grouping as this information can be arrived at in other ways. However when I cam eot make the change it became clear that this complex term is referenced in many other process terms e.g. protein-DNA complex assembly, protein-DNA complex remodeling, protein-DNA complex transport, protein-DNA-RNA complex remodeling some of which may be useful, esp the remodeling terms.

Review of Glycolysis work

OBO-Edit 2.3 release

Available at http://sourceforge.net/projects/geneontology/

Please install, test and comment. Paola has used it for everyday commits and it works fine.

Also see http://wiki.geneontology.org/index.php/Ontology_meeting_2013-11-20#Moving_logical_defs_into_the_editors_obo_file (Testing of OE catalogs, Downstream changes)

'Export' logical defs

Stemming from the latest assert report job:

ADD GO:0035928 'rRNA import into mitochondrion' GO:0019090 'mitochondrial rRNA export from mitochondrion'