Ontology meeting 2016-10-13

Attendees: David Hill, David Osumi-Sutherland, Melanie Courtot, Paola Roncaglia, Harold Drabkin, Tanya Berardini, Paul Thomas, Moni Munoz-Torres

Minutes: Moni Munoz-Torres.

Regrets: Chris Mungall

GoToMeeting invite: https://global.gotomeeting.com/join/859015101

SOP when merging terms

David H confirmed that the problem doesn’t stem from merging in OBO-Edit itself (“did a test merge in obo-edit and the resulting obo stanza looks fine. I suspect that the build problems arise in the pipeline downstream. No ghost terms after the merge and save.”), and suggests the following:

“I think that whenever we merge, we need to check the inferences file [go_inferences.obo] for references to the merged terms. If they exist, they should be deleted. Then when the terms are merged, the cross-products need to be checked and edited as appropriate. Presumably then when the write file [gene_ontology_write.obo] is committed, the ghost inferences will not exist in the inference file and the ghost terms won't be created.”

Let’s confirm. Comments here please: https://github.com/geneontology/go-ontology/issues/12729

 We’ve got the GO Build blues!
 Melanie, David, Paola, suggest that we come up with a procedure to do the merges in a better way, so that the build no longer fails. 
 If you are not importing the inferences after the merge, you’re good. 
 DavidOS suggested to obsolete the term first, but we’ve been doing this differently.  
 
 We now have a process to use OBO-Edit (you need to make sure logical definitions are consistent; remove references to the term that was merged). If you merge 2 terms that don’t have any inferences, then you’re okay - there won’t be references in the OBO file. 
 
 The instructions from David OS about Editing in Protege are in this Google doc: https://goo.gl/gHZr5Z Perhaps now is a good time to think about moving this content to the wiki.
 This replicates what we’ve always done in OBO Edit.
 We don’t lose URIs anymore. 
 Recommendation was to put old ID on the synonyms as a way of tracking.
 
 Need to come up with an SOP for making a merge on OBO-Edit as well as the one that is available on the google doc for Protege. Without being able to “test” it is a bit difficult. 
 
 *ACTION*: Paola will run through the steps with the next merge she needs to do (using OBO-Edit) - by next week. Will write up an SOP.

equivalence axioms (GCIs) required for reasoning with ChEBI missing from editors file ?

https://github.com/geneontology/go-ontology/issues/12721

gene_ontology_write.obo imports cheb_import.owl. This lacks the equivalence axioms (GCIs). But we have a file called bio_chebi.owl which has them. bio_chebi.owl was last updated 19 months ago: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/ontology/extensions/bio-chebi.owl?view=log

 During the move to using ChEBI, a few issues wrt distinctions were introduced for the purpose of reasoning in OWL for GO. We mapped equivalent axioms; some terms were conflated. In BioCHEBI, the equivalences are made, but we don’t know the downstream pipeline. 
 
 These equivalent axioms are not imported when you load gene_ontology_write.obo - they were at some point - but it needs an import statement to BioChEBI. It is in chebi_import.owl, but  nothing shows that they’re being used.
 
 We need to fix BioChEBI input, or come up with an alternate strategy.
 
 Tanya: these are being used in TermGenie. So some process is updating that file at least for TermGenie.
 Heiko once said: “ontology uses BioCHEBI”, but it is not using it now. 
 
 *ACTION*: We need Chris to please clarify.

Update: contacts for UniProt annotations

As already mentioned, we should email the GOA helpdesk at goa@ebi.ac.uk for annotations assigned specifically by EBI UniProt (GOA) curators. Where annotations by SIB UniProt curators are or may be involved, please email help@uniprot.org. They distribute updates to all UniProt curators (GOA + SIB). I added this info to the Google spreadsheet (https://docs.google.com/spreadsheets/d/1nAquF0oXHU6d638Q97ZeQ7K7DOhuru84-rOMEL4A114/edit#gid=0), but we opted not to include email contacts on the corresponding GitHub page (https://github.com/geneontology/go-ontology/blob/master/curator_group_contacts.tsv).

Also, [one for David OS?] Chris reminds us to coordinate with Seth on any metadata related to people, see https://github.com/geneontology/go-site/issues/231.

 File in GitHub does not include email addresses. 
 
 *ACTION*: GOA office will have a think about creating a @goa user for GH and get approval from GOA supervisors. 
 
 *ACTION*: For UniProt - let’s ask Sylvain about the possibility of creating a @uniprot ID on GH that can follow up as contact information. 
 
 Reminder to coordinate with Seth wrt updating people / users. We can use yaml file to find IDs.

Update: membership go-ontology mailing-list

Do we want to restrict membership of the list or should it be a public one? (The current list of subscribers can be accessed upon logging in https://mailman.stanford.edu/mailman/listinfo/go-ontology)

*ACTION*: Contact previous editors still in the list to ask them to please review whether they wish to continue receiving messages from go-ontology list.
 
*ACTION*: Contact Mike to ask him to please remove unknown address from list. (Done).

Pending TG requests

Just a note so we don't forget to deal with these:

regulation of memory terms (requested by Helen Attrill, FlyBase)

heteroreceptor complex assembly/disassembly (requested by Paul Denny, UCL)

 1. DavidOS updated us on https://github.com/geneontology/go-ontology/issues/12512 Will continue to look into it and update it.
 
 2. DavidOS, Tanya commenting on https://github.com/geneontology/go-ontology/issues/12610
 Tanya added a term, then Paul Denny requested a few additional terms. 
 
 DavidOS: we can remove the TG terms. But it depends on agreement that it’s not a good idea to add these as cell component terms, rather add some BP terms.
 DavidH: worried that curators will think that these are heterodimeric receptors. 
 “multidimeric” could also be used.
 
 *ACTION*: DavidOS will work with PaulD at UCL next week. It might be that we missed something from the paper. 
 The proposal is:
 LEGO suggestion: 'complex 1' annotate to 'receptor binding' has_input 'complex 2' with annotation extension regulates 'appropriate signal transduction' 
 
 *ACTION*: get rid of TG requests in the queue. 
 
 *ACTION*: Let’s try to model this using a LEGO model to clarify. DavidOS will start something with Paul next week.
 Main paper PMID:24157794 https://www.ncbi.nlm.nih.gov/pubmed/24157794
 Additional information: PMID: 22035699 https://www.ncbi.nlm.nih.gov/pubmed/22035699

Follow-up: GitHub projects

See background here: http://wiki.geneontology.org/index.php/Ontology_meeting_2016-10-06#GitHub_Projects

Paola added one card each to the 3 columns in https://github.com/geneontology/go-ontology/projects/1. Feel free to add as an experiment, and/or we could go over those. Then we can see how this works for us.

Melanie: https://github.com/geneontology/go-ontology/issues/12723 "hypusine pathway", discussed with David Hill and thought worthwhile to bring to editors call.

 We will look at this on the next call.
 Beginning with: https://github.com/geneontology/go-ontology/issues/12723