Annotation Conf. Call 2023-05-16

From GO Wiki
Jump to navigation Jump to search

Agenda and Minutes

Meetings and Announcements

Noctua

  • Next regularly scheduled Noctua outage:

Fall GOC Meeting

  • Virtual meeting
  • October 17th - 19th (Tuesday - Thursday)

GO Release Pipeline

  • May release coming
  • #release-pipeline Slack channel for updates

Spring GOC Meeting - Follow up

Annotation and Ontology Discussions

Annotation Review

  • Consensus that annotation reviews:
    • ensure consistency and quality of our annotation product
    • compete with other curation priorities
    • can lack intellectual reward (re-annotating older papers)
    • need to be counted in annotation metrics (see below)
  • Need to come up with new and better ways to approach annotation reviews
    • Can annotation reviews be limited to those where a 'human in the loop' is really necessary?
    • How much more can be automated?
      • ontology term obsoletions with replaced_by are automatically updated in Noctua and Protein2GO
      • what about other ontology changes?
        • for example, if BP 'phosphorylation' is being obsoleted and an annotation to MF 'kinase activity' already exists for that gene, no need to manually review
        • for example, for 'colocalizes_with' gp2term relation that we would like to obsolete, automatically migrate to 'located in' (eventually these annotations could be upgraded (is active in) or deleted over time)
  • Have different slices of GO annotations available for users
    • 'Fool proof' GAF - most precise GO terms, gp2term relations, etc.
    • 'Everything' GAF - contains all the annotations we have
    • Implications for GO analyses? display? gene descriptions?

Metrics and Attribution for Curation Work

  • Prototype of New Annotation Metrics Reports
    • Chris has worked on a script to generate new annotation metrics, based on a diff of GAF files
    • Slide from GOC meeting
    • Additional metrics we'd want to count?
      • Annotation extensions? A subset of annotation extensions, e.g. 'has input' on MFs
Strategies for Integrating GO-CAM Modeling into Curation Workflows
  • Several presentations on how different groups are strategizing to create GO-CAMs
    • Focus on specific areas of biology - use recent articles and reviews
    • Use gene sets identified by clustering of annotations
    • Use existing models as templates for seeding new models
    • Add relevant extensions during standard annotation, e.g. 'has input' for regulatory MFs
    • Think about - are the annotations for this gene what I/users would expect to see?
      • If not, can the biology of these genes be represented better in a GO-CAM?
    • Complex BP terms can now be better represented with a GO-CAM
      • BP involved in BP; BP in response to X; MF involved in BP
  • Related GO-CAM work
    • MF-to-BP relations
Use Cases for GO Annotations
  • Automated gene descriptions
  • Use GO annotations as one of the primary sources of information

Updates on Specific Areas of Ontology and Annotation

Gene Expression
  • ncRNA curation
  • chromatin remodeling and histone modifications (histone code)
Signaling Pathways
  • Is this gene in/out of the pathway? If it's out, does it regulate the pathway?
  • Signaling cassettes or modules
    • What is needed and where do they fit in to annotation practice?

Software Discussions

Noctua Updates
Centralizing GO Annotation Processing
  • David H coordinating this project
  • MGI will be a 'test' case
  • MODs do different things with their annotations
AmiGO Updates
  • Adding GO-CAM tab to AmiGO gene/protein and BP term pages (Patrick Kalita)
Ontology Editing Updates
PAINT
Large Language Models
  • How do LLMs affect GO annotation?
  • Can they help? Are they a threat?

Recommended Reading

PAINT

AmiGO

  • Patrick Kalita is working on a project to display GO-CAMs on AmiGO
    • Will add pathways widget as seen on the Alliance gene pages and in Noctua (separate workbench)

Annotation

Gene Product-to-Term Relations for External2GO Mappings

Annotation Reviews

Annotation error reports and annotation review

Error reports

Reannotating regulation of MF process terms

  • Also check annotation review tickets assigned to you/your group in go-annotation repo on github
  • Questions about either - please bring to the meeting

GO-CAM and Noctua

Ontology

Ongoing Projects

Term Merges and Obsoletions

  • New proposed SOP for handling merges and obsoletions: No longer merge terms, only obsolete. When we used to merge, we'll now do obsolete + replaced by
    • Noctua and P2GO automatically updates obsoleted terms that have a replaced_by; is this a problem for any group on the call?
    • Note that QuickGO and AmiGO also repairs these annotations during the load process
  • Streamline announcements for obsoleted terms having a replaced_by term: once per release (link to stats:
  • For consistency with the OWL representation, and with other OBO ontologies, we would like to create obsolete terms for all alt_id entries. These would have a replaced_by link to the active term. After this, the ontology would contain no alt_id tags. (https://github.com/geneontology/go-ontology/issues/24467)

Current Projects

Obsolete "Regulation of molecular function" processes terms

Attendees

  • On call: David, Debby, Dustin, Giulia, Edith, Kimberly, Leonore, Li, Malcolm, Pascale, Patrick, Petra, Raymond, Rob, Ruth, Sridhar, Stan, Steven, Suzi, Val
Retrieved from "https://wiki.geneontology.org/index.php?title=Annotation_Conf._Call_2023-05-16&oldid=85339"