5 May 2020 PAINT Conference Call

From GO Wiki
Jump to navigation Jump to search

Present

Following up on previous actions

Obsolete Uniprot entries should be filtered (ie not exported in GAFs generated by PAINT) at each PAINT release

Pascale created a ticket: https://github.com/pantherdb/fullgo_paint_update/issues/46

Alex linked the PAINT error report; is this enough to implement the filter ? --- Discussion:

Taxon constraints

The test version is available at http://panthertest3.med.usc.edu:8083/

We plan to release it before the GO meeting.

Related questions:

  1. What to do when TC is violated in the experimental annotation? Example:http://amigo.geneontology.org/amigo/gene_product/FB:FBgn0003507
  2. What to do when incorrect qualifier is used? Example: http://amigo.geneontology.org/amigo/gene_product/MGI:MGI:2137670
  3. Taxon constraint file may not be perfect.

GAF 2.2 more qualifiers

How are we going to handle these in PAINT?

https://github.com/pantherdb/fullgo_paint_update/issues/45

PAINT will "ignore" qualifiers when judging whether an experimental annotation can be used for supporting a PAINT annotation. Meaning, PAINT annotations will not require supporting experimental evidence to have same qualifiers. So for example, an exp GO annotation "Gene:00001 acts_upstream_of GO:000001" can be used to support an IBD "PTN0000001 involved_in GO:000001". And this will propagate an IBA "Gene:00001 involved_in GO:000001". The exception, of course, is for NOT qualifiers.

PAINT IBDs (thus inferred IBAs too) will use the qualifiers as below (except for NOT qualifier):

  • biological process - use "involved_in" for all annotations as default.
  • cellular component - use "is_active_in" for all annotations as default.
  • molecular function - use "enables" as default, but "contributes_to" can be used where it is appropriate.

PANTHER 15 migration in PAINT

Will do the migration

Discussion points