Annotation Conf. Call 2020-12-01

Agenda and Minutes

GAF 2.2 and GPAD/GPI 2.0

• Discussion with groups about their current status with creating GAF 2.2 files
• Note that the specifications were updated to give guidelines on what relation to use with annotations to the root nodes:
  • Biological Process -> involved_in
  • Molecular Function -> enables
  • Cellular Component -> is_active_in
• Achieving synchronicity amongst groups wrt applying gp2term relations in a GAF 2.2 file. Note that this is distinct from harmonizing on using the same relations for gp2term for the same, or similar, annotations amongst different curation groups.
• Each annotation group should be applying their gp2term relations in accordance with the GAF 2.2 file specifications and their annotation practices, as well as the guidelines for root node annotation relations.
• Alex (UniProt - GOA) and Dustin (PAINT) are joining the call today to provide more information about the annotations from their groups/pipelines as these two annotations sources are a common upstream for many annotation groups.
  • PAINT test files - ftp://ftp.pantherdb.org/downloads/paint/test/gaf2.2/
    • PAINT files use for default:
      • Biological Process -> involved_in
      • Molecular Function -> enables
      • Cellular Component -> is_active_in (for non-protein-containing complex terms)
      • Cellular Component -> part_of (for protein-containing complex terms)
    • Also use:
      • Molecular Function -> contributes_to
      • Cellular Component -> colocalizes_with
• External2GO mappings
  • InterPro
  • SPKW
  • SUBC
  • UniRule
  • EC
• Sequence-based annotation transfer and gp2term relations
  • Curators can update a gp2term relation from a less to a more specific relation upon review for a sequence-based (e.g. ISS, ISO) annotation
  • Consider a centralized reporting system when this happens to feed back to the group that made the original annotation
• Other outstanding issues/questions?
  • Working group in the new year (2021) to draft proposals for harmonizing gp2term relations for specific gene family - GO term associations

Meetings

GO-CAM Jamboree

• Held from November 16th - 30th
• ~30 people on and off throughout the week
• Focused on generating process-centric models
• Good discussions and feedback on process/workflow and tool
• REMINDER: please send Paul and/or me your best estimate on how much time it took you to create a model, or begin to create a model, with any additional thoughts on your individual experience

Meeting-less Weeks

• We are going to experiment with one meeting-less week a month (second week of the month). Any objections? :-)
• People are encouraged, though, to get in direct touch if they have pressing questions or continue to use the github trackers.

Annotation Issues

Capturing interacting taxon

• • How to capture in GO-CAM (and the MOD imports)?
  • New ontology terms and the scope of GO
• Currently, a second taxa is just captured in a separate field in the GAF (Taxon, pipe separated when more than one) or GPAD (Interacting taxon) with no relation.
• Proposal: capture dual taxon information using the annotation extension field and a 'has input' relation
• Questions:
  • Does everyone who uses the existing Taxon or Interacting taxon fields for dual taxon annotations also use annotation extensions?
    • Number from UniProt (Protein2GO):
      • UniProt 1749
      • WB 157
      • CAFA 67
      • ParkinsonsUK-UCL 28
      • ARUK-UCL 23
      • BHF-UCL 21
      • DIBU 3
      • GO_Central 1
      • MTBBASE 1
    • Other groups?
  • Would this apply to all children of interspecies interaction between organisms?

Attendance

• On call: Birgit, Bob, Colin, David, Debby, Dmitry, Dustin, Edith, Harold, Helen, Karen, Kimberly, Li, Malcolm, Midori, Pascale, Patrick, Petra, Rob, Sabrina, Seth, Stacia, Suzi, Tanya