Annotation Conf. Call November 26, 2013: Difference between revisions
Jump to navigation
Jump to search
(→Agenda) |
No edit summary |
||
Line 1: | Line 1: | ||
[[Category:Annotation Working Group]] | |||
==Agenda== | ==Agenda== | ||
Latest revision as of 16:46, 9 April 2014
Agenda
Present:
SGD: Rama, Edith, Diane
MGI: Judy, Li
Zfin: Doug
WB: Kimberley
EBI: Jane, Rachael, Alex
Pombase: Midori
Panther: Paul
Berkeley: Moni, Chris
RGD: Stan
Dicty: Petra, Pascale
TAIR:
UCL: Ruth
Jenkins email alerts
Any questions on Jenkins email alerts about your GAF files?
Curation question
- We noticed that there are some histone kinase terms that include the name of the residue that is being phosphorylated in the term name itself
(http://amigo.geneontology.org/cgi-bin/amigo/term_details?term=GO:0035174#lineage). Although histones are very well conserved, there are still some differences in the sequence and the serine is always not in the same position in all organisms, but often the context is the same (i.e it is +/- one or two residues).
- should we requesting a new term for each specific residue
- should there be a taxon constraint (because for example, position 14 in H2B may not have serine in all organisms)
Column-16 curation questions
- Forward annotation example: In yeast, CNB1 is the regulatory subunit and CNA1 and CMP2 are the catalytic subunits of the Calcineurin complex (http://www.yeastgenome.org/cgi-bin/GO/goTerm.pl?goid=5955, phosphatase complex).
DB (Col 2) | GO ID (Col 5) | ev.code | Reference (Col 6) | Extension (Col 16) |
---|---|---|---|---|
CNB1 | GO:8597 (calcium-dependent protein S/T phosphatase regulator activity) | IDA | PMID:1321337 | has_regulation_target: CNA1, CMP2 |
CNA1 | GO:4723 (calcium-dependent protein S/T phosphatase) | IDA | PMID:1321337 | requires_regulator: CNB1 |
CMP2 | GO:4723 (calcium-dependent protein S/T phosphatase) | IDA | PMID:1321337 | requires_regulator: CNB1 |
- RAD51 has increased strand exchange activity in the presence of RPA complex or RPA promotes this activity. For the RPA "promoted" part, we can't use "activated_by" relationship in col 16, because this relationship is only appropriate for chemicals. We could capture this is an forward annotation for RPA: create a new term for positive regulation of strand invasion and annotate RPA to this term with target rad51 - the problem: the subunits will have to be annotated to this term and not the complex as a whole. Can activated_by relationship be relaxed for use in these situations or do we have to wait to annotate to complexes as objects.
DB (Col 2) | GO ID (Col 5) | ev.code | Reference (Col 6) | Extension (Col 16) |
---|---|---|---|---|
RAD51 | GO:recombinase activity | IDA | PMID:8066464 | ? (Replication protein A complex serves as an accessory factor or activated_by, but RPA is a complex) |
- Would you add the substrate information in col-16 for phosphorylation BP annotation if we have already captured the substrate for the Kinase activity, MF annotation?
Discussion
- Jenkins: Looks like groups are receiving the Jenkins reports via email and are fixing their annotations. We will bring this up at Annotation calls for the next 4 months just to make sure everybody is on the same page.
- Histone kinase issue:
- Creating amino acid position specific terms in this case can lead to incorrect annotations. These terms should be evaluated and removed by the ontology group leaving the higher level term
- Curators should request PRO IDs for the modified/specific phosphorylated form of the histone and include that ID in col-16 with has_output relation.
- Calcineurin case: Col-16 annotations shown in the table are okay to have. Make an additional col-16 annotation to indicate that the three subunits are part of the complex using occurs_in (or part_of) GO:00005955.
- Chris: one should be able to visualize this data in LEGO to see how these fit.
- Ruth: For CNB1, the GO term already includes the detail about it being a regulator. In that case why do we have to say has_regulation_target in column 16? Why can't we just say has_input? Isn't that redundant? When is it appropriate to use has_input? We need more curation examples to sort this out (when to use has_input)
- RPA example: Can't be done with the current model. We need to be able to annotate to complexes directly to capture these details.
- can the limitation on the relationship 'activated_by' be lifted so it can be used to capture this experiment?
- Complexes working group is collecting the type of identifiers that groups would use to annotate. This project should be moved up in priority.
- col-16 for Phosphorylation in BP: if you have captured the substrate in col-16 for a Kinase activity term in MF, do we have to capture it again for phosphorylation in BP?
- Jenkins inferencing script can't handle col-16 yet. So, yes, capture it in BP also.
- TermGenie: Soon there will be a new feature in Termgenie to add Notes. THese notes will become public comments, so can't be crude. Note should be something like "an example of this term can be found in S. cerevisiae for gene X as shown in PMID with blah evidence. Ontology group will give a demo of this feature when it is ready.