Ontology meeting 2014-05-01
Attendees: David OS David H Tanya Jane Heiko
Minutes:
'hydrosylate' terms
Trudy has submitted to terms where the participant is 'hydrosylate' - I'm pretty sure hydrosylate isn't something that occurs natutrally - it's used in cell culture although it is defined as 'any product of hydrolysis' so strictly speaking it's in scope. To add or not?
Typo for hydolysate - as in 'some product of hydrolysis'? If so, is too broad to be referred to by a GO term. Even of it refers to some specific hydrolysate (e.g. starch hyrdosylate seems to be a common type for starch hyrolysate) it is probably still too broad as likely to be complex heterogeneous mix.
Megafile content policy re disjoints and never_in_taxon
- TODO: Disjoints to go into megafile. (subject to confirmation that we're agreed on this)
- Agree - but we need doc. DOS to put some doc in editors guide on wiki.
- TODO: never_in_taxon to be treated as an annotation in OBO - the alternative requires ad hoc extensions to OWL->OBO conversion that we should probably avoid.
* Need to confirm this and add some doc (where?) Chris decided that OBO to OWL conversion should properly roundtrip these. Heiko has implemented this - although you'll need to have the latest OWLtools and oboformat obo2owl converter installed. This changes means that never_in_taxon 'relationships' can be edited in OBO after all.
Follow-up: 'colocalizes with' nuclear chromatin
Carried over, see http://wiki.geneontology.org/index.php?title=Ontology_meeting_2014-03-13
Where are we with the planned action items? David OS made a start on this. Transitive?
The latest AI was for David OS to put together a proposal for this, and run it by Tony.
Progress:
Added to RO with URI http://purl.obolibrary.org/obo/RO_0002008
Definition: "A relation that holds between two linear structures that are approximately parallel to each other for their entire length. This relation is transitive. It i is agnostic about any parthood relationship between X and Y."
Comment: Example: if we define region of chromosome as any subdivision of a chromosome along its long axis, then we can define a region of chromosome that contains only gene x as 'chromosome region' that coincident_with some 'gene x', where the term gene X corresponds to a genomic sequence.
is_transitive: true
subPropertyOf: mereotoplogically related to
domain: continuant that 'bearer of' some linear*
range: continuant that 'bearer of' some linear* <--- Might be problematic for SO terms as range?
Skype discussion with Chris: [01/05/2014 16:58:24] Chris Mungall: I think it looks good. If you want to test edge cases then what about when regions that are distant on the linear axis come into contact and align in 3D [01/05/2014 16:58:49] Chris Mungall: The intent would be to exclude this, so I think a note in the comments is fine for that. [01/05/2014 17:10:27] David Osumi-Sutherland: "Do you have any examples in mind?" [01/05/2014 17:38:02] Chris Mungall: nope
\* PATO_0001199 - Need to update imports.
TODO: Make RO release
- Will annotation extension support the relevant targets?
- Yes: SO terms and Ensembl IDs can already be used. Rachel has given me details of what extra is needed for the relation to be used in extensions.
Follow up: editing in Protege
- First of all - could we please record here clearly what the procedure should be when loading the file in Protege. Jane and Paola thought that we were going to use the 'make' command to convert the latest editors file in SVN from obo to owl and then load that in Protege for editing (so using File -> Open), but the current gene_ontology_write.owl file in SVN has a datestamp of April 4th. Alternatively, should we File -> Open from URL -> http://purl.obolibrary.org/obo/go.owl?
Please record procedure for converting back to obo where necessary, so we're all on the same line.
Editors still don't have a standard way to convert from OWL to OBO (DS using his own bash function for this) TODO: Chris to devise simple system owl-obo megafile conversion to work in checked out repos. (Perhaps use bash script as can't have circular make.)
The current version of the plugin is available here: org.protege.oboeditor.jar
Please install and test the plugin. The installation should be simple.
- Download the jar and copy it in the Protege plugin folder.
- Start Protege, there will be a new view available
Window > Views > OBO views > OBO annotation
See here for background: http://wiki.geneontology.org/index.php/Ontology_meeting_2014-04-24#Follow-up_on_Protege
See here for wish list: http://wiki.geneontology.org/index.php/Protege_Feature_Requirements
DS and DH have experimented with this plugin a little. Seems to work well as far as it goes, although DS sees lots of extra blank lines showing up in display under synonyms.
Assert Inferences
A simplified version of the assert-inferences job is now running again. This job will only add new inferred axioms.
We had quite the back log of axioms, see here for all added axioms after the first run.
Assuming these have gone in, presumably we need to review them? DOS: Can it be run more often? TODO (Heiko): Get inference stripping back in safe form - make sure there is a ticket for this.
The previous configuration is still available as a report.
Ontology SF requests queue
SF is sort of spinning out of control again (288 open tickets as of last night). Many of these are mini-projects in themselves, but is there anything we can do to make a compromise between our limited resources and keeping the numbers a bit lower? Maybe establish a weekly SF-only day for a month (that is, on the weeks when one is not on the SF rota), and make people around us aware of that? Or shall we (and the submitters) just live with that.
Problem is that one left are often the hard ones that are blocked because of some difficulty in understanding or because of some difficult decision. Suggestion: Jamboree where each editor brings some specified number of blocked tickets to SF meeting for discussion.