Ontology meeting 2015-06-11
Attendees:
Chris, Melanie, Heiko, Paul T, David OS
Minutes: David OS
Due to limited attendance, all of the planned agenda items were pushed to next week. Instead we discussed possiblities for implementing Paul Thomas's draft molecular function overhaul.
Our previous notes on this are here: https://docs.google.com/document/d/1y-6to0vwEJHc3i3wdPnJoCAq910auPQp5jBvafpJvL0/edit#heading=h.gcg6ywbotzhz
Paul's draft in Web Protege - with comments, is here: http://webprotege.stanford.edu/#Edit:projectId=ea132f81-760a-43f2-b5a9-fbe763bb7eed
Notes on discussion:
- We discussed the possibility of defining compound functions using the pattern molecular_function that has_part some molecular_function. This DL query currently returns over 500 classes, many of the receptors and enzyme activity regulators. In subsequent discussion, it became clear that this might not be too general a grouping term as, in the refactor, Paul has: 'catlytic activity' has_part sone 'substrate binding'.
- Need to agree whether the class 'binding' should cover transient binding to substrates. Also If we are to axiomatise that substrates must be bound, we need this to be embedded in relation axiomatisation so that editors don't have to assert binding. e.g.
has_input . binds . . has_substrate
- The two branches: 'molecular component function' and molecular mechanism are NOT disjoint. e.g. the latter has 'catalytic activity' as a direct subclass, but the former subsumes many specific catalytic activities. (e.g. see the branch 'RNA structure modulator'). This means that, unless equivalent class definitions are specified for these new classes (or for existing classifications), mainting these new classifications will be hard and liable to failure.
- Many terms in the 'molecular component function branch' could probably be defined via an equivalent class axiom combining has_part MF and part_of BP clauses.
- AI: Paul T to pic 5-10 of the most important molecular component function terms and make tickets including the text definition and suggesting an axiomatisation. These should have an appropriate label.
- AI: David to start an RO ticket for 'has substrate' with draft axiomatisation
https://code.google.com/p/obo-relations/issues/detail?id=64&thanks=64&ts=1434664087
- AI: David to provide Paul T with a list of existing compound functions.
http://wiki.geneontology.org/images/9/96/MF_has_part_MF.txt