Annotation Conf. Call 2017-05-09

From GO Wiki
Jump to: navigation, search

Bluejeans URL

https://bluejeans.com/993661940

Agenda

GOC Meeting, Corvallis, OR, June 1-3rd

  • Meeting Agenda
  • Review items; link to github tickets (create a ticket if none exists)
  • Poster session

Annotation and Ontology Proposals

Modification-Dependent Protein Binding

  • Proposal from Pascale and Sylvain emailed
  • Would like to review and approve proposal on this call

The proposal is the comment near the end of issue: https://github.com/geneontology/go-ontology/issues/12787

(text starts with "Here's a version of the proposal, reviewed by the GO-editors and by @sylvainpoux"

Transcription Decision Tree

  • AI: Review annotation decision tree which is linked from the github ticket in go-annotation
  • Reminder to review the current state of the proposal on the github tracker
  • github ticket 1463

Annotation Guidelines

Review Annotation Guidelines for Cellular Component (CC) Annotations

  • GO Cellular Component annotations are meant to capture the location where a gene/gene product is active, i.e. where it executes its molecular function
  • Gene products can be found in multiple places within the cells in which they are expressed
    • Some of those locations reflect where it is believed to act, others not
      • Gene products that are transported by the secretory pathway
        • Soluble secreted proteins and transmembrane proteins can be found in a number of different cytoplasmic compartments (ER, Golgi, vesicles) as they traverse the secretory pathway
      • Gene products that are relocalized as part of the regulation of their activity
        • Internalized receptors and/or ligands can be found in endocytic vesicles
        • Transcriptional regulators can shuttle between the cytoplasm and the nucleus
  • Curators should use existing knowledge about the functions and processes of the gene/gene product to determine what is an appropriate annotation
    • For example, we don't want to say, categorically, that a receptor could never be annotated to an endocytic vesicle or a component of the secretory pathway, but if the annotation is made, it's because there is positive evidence that the receptor has a function there
  • Annotation questions
    • Is it appropriate to annotate a membrane-bound ligand expressed in a signaling cell to a location in a receiving cell?
      • ZP3 (zona pellucida glycoprotein) found in oocytes to 'outer acrosomal membrane' (of sperm cell)
      • Notch - Delta signaling is another example
    • What if you don't know whether a subcellular location is a site of molecular activity?
      • A novel protein expressed in multiple locations; mutations in the encoding gene affect known processes

Use of colocalizes_with Qualifier

 Gene products that are transiently or peripherally associated with an organelle or complex may be annotated to the relevant cellular 
 component term, using the colocalizes_with qualifier. This qualifier may also be used in cases where the resolution of an assay 
 is not accurate enough to say that the gene product is a bona fide component member. Example (from Schizosaccharomyces 
 pombe): Clp1p relocalizes from the nucleolus to the spindle and site of cell division; i.e. it is associated transiently with the 
 spindle pole body and the contractile ring (evidence from GFP fusion). Clp1p is annotated to spindle pole body; GO:0005816 and 
 contractile ring; GO:0005826, using the colocalizes_with qualifier in both cases.
  • Discussion Points
    • Assay resolution
      • A qualifier used when the level of resolution of an assay did not allow for unambiguous determination of localization
    • Peripheral association
    • Transient association
      • Gene/gene products can be found in multiple locations for a defined period of time, but if that location is where they function, curators do not need to use the colocalizes_with qualifier
        • If possible, curators can use annotation extensions to provide more information about the phase or stage during which the localization is observed
          • Example of a transcription factor that is present in the cytoplasm but then moves to the nucleus to function in response to a stimulus
      • If the gene/gene product is not believed to function in a particular location, use of colocalizes_with is not warranted
    • Use of colocalizes_with and complex annotations
    • Can we deprecate colocalizes_with? Do we now have better ways to describe what we were trying to capture with this qualifier?

Minutes

  • On call: Alice, David H., David OS, Edith, Emily, Harold, Helen, George, Giulia, Karen, Kevin, Midori, Moni, Petra, Rachael, Sabrina, Seth, Shur-Jen, Stacia, Stan, Suzi, Tanya, Val

GOC Meeting, Corvallis, OR, June 1-3rd

  • Please check the agenda, add items, if needed
  • Link all items to a github ticket

Annotation and Ontology Proposals

Modification-Dependent Protein Binding

  • New proposal from Pascale and Sylvain
  • Want to capture binding events that are *dependent* on a particular PTM, not just binding of a gene product that has a PTM
  • Annotations to existing terms (and their children) will need to be reviewed; will try not to obsolete terms and create new ones, but curators will likely need to review annotations to make sure they are consistent with the proposal
  • Previous annotations did not likely make the distinction between a PTM-dependent binding and binding a PTM protein, e.g. phosphoprotein binding
  • Will need to judge, going forward, how many PTM-specific terms will be needed
    • Curators could use annotation extensions with the has_input relation and a PRO ID for the modified form of the protein
    • However, the semantics of the AE approach would not be as clear as instantiating a term in the ontology
    • Harold added list of 12 modifications currently available in PRO to the github ticket
  • Can rename terms like 'phosphoserine binding' to 'phosphoserine residue binding' in the ontology

Transcription Decision Tree

Annotation Guidelines

Review Annotation Guidelines for Cellular Component (CC) Annotations

  • We discussed the proposal to annotate CC where genes/gene products function
  • There are some issues with this, most importantly that there are many annotations that report subcellular localization but for which we don't have information regarding the function of the GPs at that site. What do we do with those annotations?
  • Options:
    • Leave as is, since the (implied) part_of relation is fine, i.e. using part_of we're not saying that the activity occurs_in that location
    • Adopt a new CC qualifier, e.g. is_active_in, to use for those CC annotations where we know the gene product is active
  • We need to discuss more specific curation cases where there are differences in annotation practice to see if we can refine our annotation guidelines/principles. For example:
    • Annotation of gene products that traverse the secretory pathway en route to the location where they are active, e.g. plasma membrane, extracellular space
    • Annotation of transcription factors that shuttle between the cytoplasm and the nucleus
    • Annotation of membrane-bound ligands that signal to receptors in an adjacent cell

Next Annotation Call - May 23rd

  • Discussion of Noctua models
  • Petra will present a model
  • Still have one other time slot open for another model presentation