Ontology meeting 2013-09-26

From GO Wiki
Jump to navigation Jump to search

Attendees:

Minutes:

Bumped from last week: Equivalent classes for Acid/Base transport and transport activity

Background here: http://wiki.geneontology.org/index.php/Ontology_meeting_2013-09-05#Equivalent_classes_for_Acid.2FBase_transport_and_transport_activity_.28bumped_from_last_week.29

Discussion:

  • Q. Is ammonia transporter the same as ammonium transporter? Ammonia is a gas, does it even exist in vivo?

Harold: Some controversey about the state of ammonia/NH4 while it crosses the membrane. Is it NH3 while crossing?
David H.: If it is NH4 before and after transport, then it's NH4 transport, we don't need to worry about mechanism. - seems to be general agreement to this.

  • Silicates:

Muliple silicates - which get merged with salic acid?
Harold: Merge salic acid transport with silicate(O2).
David H emphasises again: It's the start and end products we care about. Not mechanisms/intermediates

TODO make a ticket merge transporter activity for ammonia/ammonium & salicate/salicic acid (O2 form), hexuronic acid/hexuronate

New OBO-Edit release

There is a new beta release for OBO-Edit available (OE 2.3-b6). It contains bug fixes for the merge behavior (issues bugs/1191 and bugs/1187) and a fix for the DAG4DAG plugin on Java 7.

The release is available in the usual place on the GeneoOntology SourceForge site:

http://sourceforge.net/projects/geneontology

Have we all downloaded/tested it? Any feedback?

Discussion

A number of editors have used and some have made commits with this version. No problems found.

Review Meeting Schedule

Follow-up from last week: DOCUMENTATION FOR NEW WEBSITE

Copied over from last week:

  • Who is dividing up the old pages to distribute to people to rewrite?
  • Who is rewriting what?

Chris: Moni has been doing some excellent work on it, we should all look at it, but the pages are not easy to access now.

AI from last week: Chris to check with Moni and point us to the right links; then we'll come back to this.

Succinct term definitions

Stemming from Chris's email. Should we keep the genus folded as general policy?

Discussion

Chris' email: "When I read GO definitions, I often have a hard time finding the differentia - it's tacked on the end of a large phrase that is repeated in the parents. I hypothesize - with zero empirical evidence to back me up - that a more succinct style would lead annotators to better select the most precise term with greatest accuracy. The succinct style would only lose redundant information that is repeated in the parent, so there would be no actual information loss.

With the right tooling support it should be possible to allow a succinct definition to be unfolded mouse click by mouse click, which allows the best of both worlds."

General agreement that genus should be stated more clearly, but also some worries (David OS, Jane) that simply naming the genus term without further explanation may not be clear enough for users/annotators in some cases and may lead to stilted definitions (cf patterns for development/differentiation).

David OS: Can't expect annotators in a hurry to read parent term definitions too.
Jane: "There's a reason we have humans editing these things"

Chris: Perhaps best to use a pattern where the genus is just stated in the first sentence and is unfolded, if necessary, after the first sentence.

General agreement that Chris' suggestion is good.

TG template for 'import into'

We get several requests for 'chemical X import into cell/cellular component' (see https://sourceforge.net/p/geneontology/ontology-requests/10395/). Can't remember if we discussed having such templates or not? Searching Jira or wiki didn't help much...

Pattern for linking catalytic activity terms to CHEBI

Apologies if this is a noob question and so the wrong place to ask.

DL query: 'catalytic activity' that 'has participant' some 'chemical entity' => only 1 subclass (and this is through inheritance from a binding term), but very many of the classes could/should reference CHEBI. and there is a serious need for auto-classification of catalytic activity terms.

From the 'Dovetailing biology and chemistry' paper I get that this is meant to work via referencing Rhea, which in turn references CHEBI. But as Rhea doesn't seem to classify chemical reactions (e.g. have a term for phosphatase reactions), how will this give use the auto-classification we need? Any scope for using the classification implicit in EC numbers (phosphatases being everything in 3.1.3.n)?

Discussion

Links to Rhea are here: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/scratch/rhea/
Using xrefs - at least for now. Includes mapping to EC numbers: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/scratch/rhea/go-rhea-align-via-xref.rpt?revision=10142&view=markup

Rhea will give us some of the leaf node definitions and links to chebi - but we will need to classify ourselves - at least partly based on EC. Also - rhea don't have all the terms we need. They have added ad hocbefore. Planning to request wholesale addition.

Chris suggests heuristic - if 1:1 mapping between Rhea and EC we use EC.

David OS: Once we've got catalytic activity functions defined and classified, it would be nice to use function to automate classification of cognate process term (phosphatase/dephosphorylation). This may require a reverse of the direction we're using now though (process has_part function, rather than just function part_of process).
Chris: Agree we should aim to automate this.