08 FEB 2011 RefGen Phone Conference (Archived)

From GO Wiki
Jump to: navigation, search

Present

Pascale, Mike, Paul, David, Kimberly, Brenley, Li, Val, Mary, Brenley

Action items

  • Discuss the possibility of running a script for triangulation of annotations with GO managers

PAINT curation: Examples where going beyond triangulation is justified

  • PTHR12733 (we did last time) [[1]]
  • PTHR11568
    • not discussed today but a very nice example from Rama - sequence is extremely conserved, annotation to AN0 can be done even though that's beyond triangulation.


Discussion on PAINT curation process

  • Are PAINT annotations ISS, or should it be something else? (David)
    • PAINT annotations are based on phylogenetic trees, which a based on sequence information. It could be considered a child of ISS. There was much concern that the PAINT annotations not all triangulation, so seemingly there are two methods. Pascale : The PAINT annotations are really done to an ancestor sequence, and then the descendants inherit those annotations. The curation process is very different from other ISS methods, but not more different than BLAST and InterPro differ from each other. Also, for other ISS methods, there is no trangulation involved: it's always a pairwise relationship, which it is also the case when doing PAINT curation, between a descendant and its ancestor, and then from an ancestor to its descendants.
  • Should we run a script to precompute the 'triangulatable' annotations ? (Li)
    • Pascale: yes we can do that, but there is still value to do the manual annotations.
  • Should PAINT 'fast pass curation' involve reading the literature (Li)
    • Pascale: yes, it's in the SOPs - I always read one review and/or the wiki gene page. It is a bit of an investment (15-30 minutes), but it saves a lot of time to do the curation, especially where there are a lot of annotations. It helps focus on the important annotations.


  • We should do some stats to compare triangulation versus more extensive PAINT curation (Li)
    • Pascale: yes definitely, that's why we wanted do get a number of trees curated, in order to do that. Right now we have 33, it'd be nice to have more.
  • Kimberly, David: There is an understanding outside the restricted group of PAINT curators that PAINT is only doing triangulation. It should be made clear that this is not the case.
    • Action item? Pascale: Not sure how to do this. We're writing papers. GO_REF:0000033 does not mention anything about triangulation:

  • go_ref_id: GO_REF:0000033
  • title: Annotation inferences using phylogenetic trees
  • authors: Pascale Gaudet, Michael Livstone, Paul Thomas, The Reference Genome Project
  • year: 2010
  • abstract: The goal of the GO Reference Genome Project, described in PMID 19578431, is to provide accurate, complete and consistent GO annotations for all genes in twelve model organism genomes.To this end, GO curators are annotating evolutionary trees from the PANTHER database with GO terms describing molecular function, biological process and cellular component. GO terms based on experimental data from the scientific literature are used to annotate ancestral genes in the phylogenetic tree by sequence similarity (ISS), and unannotated descendants of these ancestral genes are inferred to have inherited these same GO annotations by descent. The annotations are done using a tool called PAINT (Phylogenetic Annotation and INference Tool).

url: http://gocwiki.geneontology.org/index.php/PAINT