LEGO May 4 2015
Continued Discussion of March Consistency Exercise Paper
- Here are the links to the models built by David, Kimberly and Rama
- Use of in_absence_of and in_presence_of annotation extensions.
- How to adequately capture the effects of protein binding on subsequent complex formation and resulting molecular functions.
Begin Discussion of April Consistency Exercise Paper
- Use of occurs_in relation for molecular function annotations
- What constitutes sufficient evidence for adding this extension, as CC annotations may be from a different experiment than the MF annotation?
- For example, protein binding based on Y2H or whole cell lysates (IPI) with separate experiments describing localization (IDA).
- Links to models built by David, Kimberly and Rama
- In attendance: David, Heiko, Kimberly, Paul T., Rama
- We all looked at the model Paul built.
- We want the ability to capture functions of complexes. Noctua should load complex IDs from IntAct. This will allow curators to accurately describe that a complex executes a molecular function, rather than having to say a molecular function occurs_in a complex, which is not correct. Since complexes are included in Cellular Component, should we make this clear on the wiki documentation for occurs_in? http://wiki.geneontology.org/index.php/Occurs_in
- Adding evidence - double-click to add.
- Most curators also had an annotation to Bfa1 for GTPase inhibitor activity. Should this be kept in the LEGO model but not necessarily connected to the downstream process, i.e., mitotic exit?
- Following on from the GTPase inhibitor activity annotations, we talked about the extent to which curators use the Discussion section of papers to inform GO annotations. Probably most curators do not spend much time on the Discussion, but rather curate directly from the experiments described in Results. Would this be different for LEGO curation? Will some experimentally derived annotations be included as independent GO annotations in the GAF, say, but left out of the LEGO model?
- Would like to start adding Noctua feature requests to Github. We know that it may be some time before these requests can be addressed, but we don't want to lose track of them and also, some of these requests could have consequences for current development.
- Thinking ahead - what are plans for training other curators in LEGO curation?
- Make a github ticket for adding IntAct protein complexes to Noctua.
- Heiko will let Seth know that curators may add feature requests, but these are not immediate needs.
- David, Kimberly, and Rama will review their March paper models to compare with Paul's version.
- Continue to work on LEGO model for the April consistency exercise - http://www.ncbi.nlm.nih.gov/pubmed/24349431