Annotation Conf. Call, March 24, 2015
Slides from last call, filtering script
- I have loaded the slides from last call (GAF pipeline): http://wiki.geneontology.org/index.php/Annotation_Conf._Call,_March_10,_2015
- Have added detailed instructions on running the filtering script yourself - http://wiki.geneontology.org/index.php/Annotation_Advocacy_and_Coordination#How_To.3F
Annotation consistency exercise
- SGD has selected the following paper for the exercise.
This is a biochemistry paper and the focus is on 3 genes: TEM1, BFA1, BUB2. There are some col-16 data in this paper.
- If any of you want another 'How to' question answered please bring it up and we will be happy to post detailed instructions.
- Next call on April 14 is cancelled. We will meet on April 28th.
- Good news is we were 100% consistent in picking the right GO terms and evidence codes. This paper was mainly chosen to sort out col-16 inconsistencies
- MF annotation to TEM1 was spot on from everyone
- MF annotation for BFA1:
- this protein is an inhibitor when present alone and is an activator when Bub2 is present. How do we express that detail in col-16. Capturing the inhibitor annotation without col-16 would be almost wrong.
- would an alternative annotation approach be to annotate Bfa1 to the GTPase inhibitor activity term and the Bfa1-Bub2 complex to the GTPase activator activity term?
- we agreed that the inhibitor annotation should get 'has_regulation_target: TEM1' and in_the_absence_of: BUB2' in col-16. The ontology editors will discuss the use of in_the_absence_ of and in_the_presence_of at their next call. In light of this paper these relationships should not be obsoleted.
- we need to also consider the identifier used in Col. 16. If the regulation target is a gene, transcript, protein (or something else), we should use the appropriate identifier to indicate which entity is being regulated.
- Doug asked if it makes sense to capture the 'in_the_absence_of' part since there are so many things that can be absent (e.g. tubulin is not present in the assay). We can't capture a NOT annotation for everything that a protein doesn't do. But in this case the authors specifically showed the role of Bub2 and it is okay to capture this detail. This would be a good case for LEGO curation.
- Aleks: Would you capture the GTP binding and GDP binding since the authors made a big point about it (measured kd etc). Normally one wouldn't because it is implicit in the term definition, but in this case it is okay.
- BFA1 contributes_to activator activity in the presence of Bub2: 'has_regulation_target: TEM1' and in_the_presence_of: BUB2' in col-16. The contributes-to qualifier requires that the proteins are part of a complex. So make sure the two proteins are annotated to the complex
- Should we annotate to the higher leverl GTPAase activator complex or the specific Bfa1-bub2 complex (which is very yeast specific). The Bfa1-bub2 complex was added by IntAct curators. If pombe and other groups need to use this term then the definition should be generalized.
- should we make the BP annotations? won't they inferred from MF-BP inferences? yes they should be. But it is not clear if col-16 data are also transferred. Rama will check with Chris.
- we should also request a new term 'GTP dissociation factor' to annotate BFA1
- NAS and TAS: For the most part we are avoiding author statements.
- Summary of annotations from this paper is available in the google doc (link above)