Manager Call 2016-07-20: Difference between revisions

From GO Wiki
Jump to navigation Jump to search
No edit summary
 
Line 15: Line 15:
==NAR paper: GO update==
==NAR paper: GO update==
* put ideas here: https://docs.google.com/document/d/1K0oG6rDWS_OxNnecdcXRh-ChECAwco48RPXtlJfNwQo/edit?usp=sharing
* put ideas here: https://docs.google.com/document/d/1K0oG6rDWS_OxNnecdcXRh-ChECAwco48RPXtlJfNwQo/edit?usp=sharing
=Minutes=
== ID space in GO annotations ==
in progress, to be kept open and reconvene at next call.
David H: each coding MGI has a UniProt ID in the dbxref column. Would be easy to mine this field. Was to do with PAINT and loading GPI file (follow-up on LEGO call discussion from Suzi)
'''AI: David H to add this in the documentation for the GPI''': Any gene coding should have a dbxref to UniProt
== Protein complexes ==
Pascale: when she PAINTs she doesn’t use a lot of contributes_to. However when you don’t know much about complex you would annotate all subunits, but if you know you would annotate only the right one, which seems uneven -> sometimes more information if we know less.
Paul: said he would write something up. Need something to formalise the propagation rule, difference if multiple subunits are responsible for function but not all of them in the complex, vs we don’t know which ones are involved and we propagate to all. If complex has function that is not found in specific subunit, then we want to use contributes_to as some gene products do contributes to the complex activity. Then other case where you just don’t know, and annotation was made to complex as we don’t know which subunits are involved: in this case we also used contributes_to but maybe we should have a weaker qualifier, e.g., may_contribute_to?
In the case where we don’t know, how do we derive gene level annotation from complex level annotation?
=> we need some way to distinguish between cases: compound activity from complex, specific activity from specific subunits, or we don’t know.
=> '''AI: Paul will add text to [[Macromolecular_complex_annotation_July_2016]]'''
== Noctua workshops ==
- doodle poll for UK workshop, only handful of dates seem to work.
- should we add a day to USC workshop? David H: we need to reach curators who may not attend GOC meeting
- get people to register for USC meeting to gauge target audience. Kimberly to add to logistics of meeting “would you attend a Noctua workshop before or after?” - Done
== invitation for NAR update==
Paul started google doc. Deadline September 15th but we probably have a few weeks after that, so October 1st.
AI for all to go through what we said last time and what we want to emphasise as being done since then. '''DL: August 15th for adding ideas/create outline.'''


[[Category:GO Managers Meetings ]]
[[Category:GO Managers Meetings ]]

Latest revision as of 11:57, 20 July 2016

Agenda

Last meeting

http://wiki.geneontology.org/index.php/Manager_Call_2016-07-06 Were the two agenda items resolved ?

  • Identifier space in GO annotations (GAF, GPAD)
  • Protein Complexes

Protein Complex annotation guidelines (proposed)

Noctua Workshops

  • Add extra days at USC meeting for a Noctua workshop?
  • Have two workshops in the UK? It seems like the number of people wanting to attend is going to be large.

NAR paper: GO update


Minutes

ID space in GO annotations

in progress, to be kept open and reconvene at next call. David H: each coding MGI has a UniProt ID in the dbxref column. Would be easy to mine this field. Was to do with PAINT and loading GPI file (follow-up on LEGO call discussion from Suzi) AI: David H to add this in the documentation for the GPI: Any gene coding should have a dbxref to UniProt

Protein complexes

Pascale: when she PAINTs she doesn’t use a lot of contributes_to. However when you don’t know much about complex you would annotate all subunits, but if you know you would annotate only the right one, which seems uneven -> sometimes more information if we know less. Paul: said he would write something up. Need something to formalise the propagation rule, difference if multiple subunits are responsible for function but not all of them in the complex, vs we don’t know which ones are involved and we propagate to all. If complex has function that is not found in specific subunit, then we want to use contributes_to as some gene products do contributes to the complex activity. Then other case where you just don’t know, and annotation was made to complex as we don’t know which subunits are involved: in this case we also used contributes_to but maybe we should have a weaker qualifier, e.g., may_contribute_to?

In the case where we don’t know, how do we derive gene level annotation from complex level annotation?

=> we need some way to distinguish between cases: compound activity from complex, specific activity from specific subunits, or we don’t know. => AI: Paul will add text to Macromolecular_complex_annotation_July_2016

Noctua workshops

- doodle poll for UK workshop, only handful of dates seem to work. - should we add a day to USC workshop? David H: we need to reach curators who may not attend GOC meeting - get people to register for USC meeting to gauge target audience. Kimberly to add to logistics of meeting “would you attend a Noctua workshop before or after?” - Done

invitation for NAR update

Paul started google doc. Deadline September 15th but we probably have a few weeks after that, so October 1st. AI for all to go through what we said last time and what we want to emphasise as being done since then. DL: August 15th for adding ideas/create outline.