Annotation Conf. Call, January 28, 2014
Enzyme regulator activity (GO:0030234)
- At the last call we agreed that annotations to this and its children terms be limited to gene products that directly regulate (e.g. bind to) an enzyme. But what are the sorts of experiments that one would see to show direct binding and regulation?
- Does the binding and regulation need to be demonstrated in the same experiment? Is this even possible?
- If the answer to #1 is no, is it sufficient to use an IDA evidence code if the binding is demonstrated in, and annotated from, another experiment, either within the same paper or in another paper?
- If regulation is demonstrated by an in vitro assay, but binding is not demonstrated, what should the annotation be?
- If binding is demonstrated in another organism, is it okay to infer binding based on sequence similarity and make annotation for regulator activity and an ISS annotation for binding?
- Should the term's parentage be changed in the ontology to be a child of enzyme binding (the definition has already been changed to include the binding component)
- How do we focus our efforts on re-annotation if we change how this term should be used?
Annotation of targets of Processes (Ruth)
Reactions break down collagen. Others have applied GO:0030574 “collagen catabolic process” to enzymes that catalyze these reactions; Reactome has also applied it to substrates and products of those reactions to annotate our pathway REACT_150401 “Collagen degradation”. That's a clear difference. But can a reaction really be described only in terms of its catalysts without regard to its inputs and outputs? We agree that P02462 is a target but would argue that the word describes its role in the process.
Future Annotation calls
we would like to have an open discussion on what are the sorts of issues the Annotation call should focus on.
SGD: Rama, Diane
Berkeley: Heiko, Chris, Moni
MGI: Judy, David Hill, Li
UCL: Ruth, Ana, Paul
EBI: Paola, Jane, Susan Aleks, Rachael
Panther: Paul T
Enzyme regulator activity
Ruth: If the binding evidence is not in the same paper, and we decide to combine two pieces using IC, then we will end up having lot of ICs and kind of hide the experiment evidence.
Val: If authors don't show binding, then should the annotation be to BP term?
David: Author performs invitro assay where he purifies enzyme, measures activity, throws in the regulator and shows up/down activity, and concludes it is a direct regulator. What do we do in this case? Author did not show direct binding.
Kimberley: Since the definition of the term includes the binding component, it limits our interpretation.
David: biochemical enzymes can be regulated allosterically or by modification. How do you capture that kind of regulation? Does any enzyme that modifies the protein also get enzyme binding?
David: is the term definition too strict or our annotation paradigm too strict. Latter!
Curators should assume that there is binding if the author says so (curator doesn't have to hunt down the binding experiment).
Rama: aren't we being inconsistent here? When we reannotated the Txn factors, we required both DNA binding (or protein binding) and transcription regulation to annotate to an appropriate MF term (i.e. if the authors showed 'how' the protein accomplished its role).
David: in TXN, we had specific terms to capture the mechanism (protein binding vs DNA binding). May be we should make enzyme regulator terms that captures the mechanism of regulation (allosteric vs protein modification)
Bottomline: Curator should use their judgement when interpreting the data (is there a context provided for the experiment).
Post Meeting: Val sent an improved definition:
Modulates the activity of an enzyme by allostery, direct modification, blocking the catalytic site or any mechanism which involves a direct interaction with the enzyme. An enzyme regulator activity changes the intrinsic catalytic activity of the enzyme toward its substrate, independently of substrate availability.
annotating to targets of processes
Peter agreed that they annotate differently compared to GO.