Ontology meeting 2021-02-22
- 1 Attendees
- 2 Agenda
- 2.1 Announcements
- 2.2 Discussion
- 2.3 xref types
- 2.4 Other Business
- Group members: Pascale, Karen, Harold, Raymond, Peter, Jim, Tanya, Kimberly, David, Chris, Paul, Eric
- Present: Pascale, Karen, Harold, Raymond, David, Kimberly, Jim, Chris, Peter,
http://wiki.geneontology.org/index.php/Guidelines_for_using_ChEBI_terms This is linked from the main ontology documentation page: Ontology_Development
ShEx vs ontology relations
Should they align ? See this ticket: https://github.com/geneontology/go-ontology/issues/12349
Not all children of GO:0015318 (inorganic molecular entity transmembrane transporter activity) are inorganic
Should we merge this ticket and get rid of some grouping classes ? https://github.com/geneontology/go-ontology/issues/20872#issuecomment-780882533
Cellular Component, ShEx, and GO-CAMs -- BLOCKING FOR GO-CAM
- The ShEx specifications state that MFs can occur in an Anatomical Entity and that GPs or Protein-Containing Complexes can be located in an Anatomical Entity.
- Most non-protein-containing complex GO CC terms are a child of Cellular Anatomical Entity which is a subclass of Anatomical Entity.
- However, annotations to root Cellular Component result in ShEx violations because root Cellular Component is NOT a subclass of Anatomical Entity.
- How do we fix this?
- Change root Cellular Component annotations to Cellular Anatomical Entity (throughout GO)
- Change ShEx value to an OR statement, for example: MF occurs_in AnatomicalEntity OR CellularComponent
- Work on RO to see if there is a relation that could be used as a parent for all children of Cellular Component relations - overlaps?
- Take this issue to the RO call(s)?
- Move Protein-Containing Complex out of Cellular Component into its own GO aspect and make Cellular Component a subclass of Anatomical Entity
- Another option (pascale) would be to use 'Cellular Anatomical Entity' and the PMID you are curating; and filter that from the exports. Only Cellular_component and ND and the GOREF for ND do get exported; otherwise ND annotations get filtered out.
- Chris: suggests keeping the current practice and use root Cellular Component and ND.
- Action would be to change the ShEX to allow a direct annotation to occurs_is/is active in root Cellular Component
- Documentation: Guidelines_for_'database_cross_references'_and_'definition_database_cross_references'#Adding_cross-reference_types