Ontology meeting 2013-03-21: Difference between revisions

From GO Wiki
Jump to navigation Jump to search
Line 51: Line 51:
**Sometimes for electronic annotations, mappings to terms in this subset are legitimate e.g. response to stress. We think we may need to split this set into 'do not use for manual annotation' and 'do not use for any annotation'. The latter would be for e.g. the x-part terms.
**Sometimes for electronic annotations, mappings to terms in this subset are legitimate e.g. response to stress. We think we may need to split this set into 'do not use for manual annotation' and 'do not use for any annotation'. The latter would be for e.g. the x-part terms.


<font color="purple">All: sounds reasonable. And we'll reduce the stringency of that check, so the errors will go.
ACTION ITEM: Jane will mark the ones for the 'manual annotation' set, and will let Heiko know when she's done.</font color>


===Web Presence===
===Web Presence===

Revision as of 14:55, 21 March 2013

MINUTES: Paola

ATTENDEES: Paola, Chris, David, Harold, Heiko, Tanya, Jane


Cell cycle

1) Devote some time to bring Chris up to speed on the cell cycle meeting at EBI

2) Make a decision on cell cycle phases

3) Other follow-ups from last week: http://wiki.geneontology.org/index.php/Ontology_meeting_2013-03-14#Cell_cycle


David summarizes the two proposals that were made at the meeting.

Chris: step one: make a parent class called 'phase' under 'biological process' and make it disjoint, put it in x-disjoints, do it in Protege as we did the other disjoints.

David: but then the current def. of BP doesn't include 'phase'.

Chris: then we'd need a separate ontology, but we could do it in steps, and not worry about that. Eventually, we might move to the 'during' relationship.

Jane: FlyBase uses the 'phase' terms (not fully correctly) for expression and phenotypes, so they wouldn't want the terms to go altogether...

All: we'll go with Chris' suggested strategy, but first we'll give curators notice so they have time to move their annotations.

ACTION ITEM: Jane will email round, giving a deadline of end of April.


Annotation relations

We've been discussing with Tony and Rachael ways we might constraint the number of relationships available to an annotator making an annotation extension. Tony would rather not have protein2go have to reason within the session for speed reasons...could we make subsets based on the object type in c16 e.g. cell type, gene, protein?


Chris, about subsets being the right way to do it: agrees, it wouldn't be limiting. So, yes, as long as there are constraints on the relationships. There could be a suggested set that would work 90% of the time. David suggests using categories of GO terms... but there's loads of different terms. We'll think about that. So we could do 3 sets: common, advanced, and complex. We could do that by reviewing the existing annotations.

David: in many cases, current annotations will have to be made more specific.

ACTION ITEM: We'll discuss further in one of the break-out sessions at the GOC meeting, on the Friday afternoon.


(also message to Chris from Tony - have you done anything about the domain and range constraints yet? If not, do it :-)) He'll be starting work on modifying their parser next week.)


Yes he'll do it.

interpro2go errors

  • I spoke with Tony and Rachael about the bad annotations being propagated from interpro2go mappings to terms in the 'do not annotate' anti-slim. Two conclusions:
    • They're quite keen to start including this subset in their DB (it's not there right now) which would catch these upstream
    • Sometimes for electronic annotations, mappings to terms in this subset are legitimate e.g. response to stress. We think we may need to split this set into 'do not use for manual annotation' and 'do not use for any annotation'. The latter would be for e.g. the x-part terms.


All: sounds reasonable. And we'll reduce the stringency of that check, so the errors will go.

ACTION ITEM: Jane will mark the ones for the 'manual annotation' set, and will let Heiko know when she's done.

Web Presence

David to report on (copying over from last week) making a shared doc with proposed new upper structure for a/ Main web documentation b/ Internal documentation in the Drupal book style. We can then all edit and migrate existing web content across.


SF upgrade

Follow-up from here: http://wiki.geneontology.org/index.php/Ontology_meeting_2013-03-7#SF_upgrade

Jane to report on migration of categories.


Protein complexes

Harold to follow up on checking the protein complex xps before we incorporate (see http://wiki.geneontology.org/index.php/Ontology_meeting_2013-03-7#Follow-up:_Protein_complexes).