25 FEB 2014 PAINT Phone Conference call

From GO Wiki
Jump to: navigation, search

Present: Pascale, Moni, Karen, Li

Discussion topics:

New version of PAINT (beta73)

  • Fixes the problem with the NOT (can propagate any term as long as there is a positive annotation)
  • It also fixes the problem with the labeling of the matrix - Suzi had made it such that parents of an annotated term were labeled like that annotated term, which made it very confusing to find what term a column actually corresponded to.
  • One remaining bug (reported by Pascale Feb 21, 2014, email to go-discuss 'wrong Dicty annotation in PTN000012953 family'):
    • Issue: There is a not annotation to a term X, and I am propagating an annotation to a parent term. I think the protein with the NOT annotation could inherit the annotation to the less specific term. Specifically we have not 'extracellular ATP-gated cation channel activity' on the dicty sequences and I am propagating 'ATP-gated ion channel activity'. It seems reasonable that the proteins with NOT 'extracellular ATP-gated cation channel activity' could still have ATP-gated ion channel activity' (note: family is PTHR10125)
      - Not discussed during the meeting, but related. This seems very similar to an issue Karen reported in late 2013. In some RNA pol subunit families, I have blocked propagation of a specific term, e.g. "RNA polymerase II activity", to a granular node and then propagation of a more general term like "DNA-dependent RNA polymerase activity" also gets blocked. What I would like to happen is for the general term which is valid for the entire family to propagate to sequences where I have blocked propagation of a more specific term.

DNA repair/DNA replication project

  • See http://goo.gl/BaQxMC
  • Families were obtained as follows: (1) Searched for DNA repair in AmiGO2. (2) Took all 'is_a' children terms, and look for the Panther families containing proteins with experimental annotations
  1. Missing families Karen pointed out that there are many families missing, at least for the DNA polymerase subunits.
    • Discussion: (Pascale): The reason for selecting those is that otherwise we cannot do the most important annotations in PAINT if the primary annotations are not there.
      • We agreed to add missing families to the list, with a note that they cannot be annotated yet. If time allows we'll also communicate the information to the relevant MOD so that the major function of those proteins would be annotated.
  2. Ontology development and DNA repair Jane wrote to GO managers asking that we coordinate better the PAINT projects with the ontology group if possible. She is concerned about the PAINT project on DNA repair because there are a lot of outstanding SF items.
    • Discussion: (Karen): Eurie Hong had requested/created a lot of terms for DNA repair. (Pascale) Also, we never encountered major difficulties with DNA repair annotations. Decision: We agreed to go ahead and start PAINTing the DNA repair families. If we have too many problems we'll revisit the decision.

GOC meeting

- Pascale posted a project report of the Apoptosis annotation project on the corresponding wiki page: http://wiki.geneontology.org/index.php/PAINT_-_Apoptosis - This is the basis for what we'll present at the GOC meeting. Take home message to communicate: the most useful lesson learned from the apoptosis project is that it's a great example to illustrate the difficulty to annotate processes versus readouts. - Slides will be sent soon for feedback.