Annotation consistency: x protein binding and with
(anyone can add themselves)
- Becky Foulger, David Hill, Ruth Lovering
This annotation query was discussed on the Reference Genomes conference call on Tuesday. From that it's clear that different groups use different criteria for when they use children of protein binding terms. In a protein interaction between X and Y, some groups annotate protein X to 'transcription factor binding' if protein Y was thought to be a transcription factor based only on sequence similarity or domain structure. Other groups are more stringent and require experimental evidence that protein Y is a transcription factor.
One point that was brought up is that for a protein A to be assigned a role in a particular process, through association with protein B, protein B must be experimentally shown to have a role in this process. Therefore, is this akin to this particular curation issue?
The discussion moved on to whether it mattered that different groups have different criteria.
So..... it was decided to add this curation issue to the agenda for the next consortium meeting. Before this, it would be useful if each group could state their current annotation policy on annotating children of protein binding terms- ie. whether your group requires experimental confirmation for the function of protein Y.
[taken from Becky Foulger's email to the GO list on the 15-08-08]
2. Proposed solution(s)
3. Comments/counter arguments
Other discussions of relevance: