Ontology meeting 2014-02-06
Attendees:David OS, Chris, Tanya, Jane, Paola, Heiko, Harold, David H
Minutes: David H
Fixing assert inferences cycle
Any Progress?
Still on hold for now. Still a bug in the report. Open a JIRA item to report this (David OS). For the moment we are nor not asserting any inferences. Should we turn them on? Remove the redundancy stripping and then turn them back on.
Response to acid as a stress
Stemming from https://sourceforge.net/p/geneontology/ontology-requests/10492/
We discussed this on a previous Editors call. We thought a good strategy would be for MGI to look at their manual, direct annotations to 'response to acid' and see if they can be moved elsewhere. Based on the outcome, contact other databases if necessary so they also look at their annotations. Then decide what changes we may want to make in the ontology. Where are we with this please?
Most of the annotations in MGI were ok. David and Harold to go back to the SF item and add comments.
CL TermGenie
Is now available for basic cell types defined by parthood. We can add the GO eds to the permissions table - who else from GO?
Parthood puts them in the context of an anatomical structure. Probably the people who are doing annotation extensions will want these. How will this be integrated to protein2go? All MGI curators will need access.
Documentation for megafile
Proposed import dependencies and dataflow post-megafile:
Autogenerated example: http://purl.obolibrary.org/obo/go/extensions/go-plus
Everyone still needs to do some things with GO-plus. Load the file--save--convert to owl etc. Check in the same time next week. If all is good, then we will make the switch after the call. SOP for editing will not change. We will no longer edit logical defs in a separate file.
Follow-up: Import vs transmembrane transport
Please add your comments to Paola's proposal(s) here: https://sourceforge.net/p/geneontology/ontology-requests/10621/?page=1
Can we think of any other ions? Yes. If we made the link to the transmembrane transport, would we create misannotation? If we make the child, then the annotation will not be specific enough. We are just talking about the high-level terms here. If we need to look at annotations anyways, then why not create a new term and migrate annotations.
Protozoan classes
https://sourceforge.net/p/geneontology/ontology-requests/10438/
Protozoa is a historical/convenience grouping that includes many unrelated taxonomic groups. We previously discussed whether we should get rid of classes that refer to protozoa altogether - or try to automate by making a taxon union class. We have since been in touch with Thomas Mock, who sent this excellent page summarising problems and detailing potential solutions: http://www.iaszoology.com/protozoa-classification/
I've had a go at mapping the 14(!) phyla of the MODIFIED SLEIGH’S SYSTEM (by A. Pechenik, 2002) to NCBI. Some phyla map directly - others map to multiple groups.
It's looking to me like constructing and maintaining a union class is too hard. So perhaps we should make classes for the major phyla from Sleigh's system for which we have relevant annotations? We could then add 'response to protozoan' as a broad synonym to each.
Comments from Chris: I favor ditching this basket grouping, or at least discouraging future annotations to it. Picked an annotation at random, Slc11a1, http://www.ncbi.nlm.nih.gov/pubmed/8033407, it should be to the more specific T. gondii.
It looks very messy trying to make a union class for protozoa. It is not realistic to make a union class with some of the examples. Maybe we just leave those terms (response to protozoa) there as leaf terms.
Protein complexes with other components
Under what circumstances should we allow the addition of terms for protein complexes in combination with some other factor? Presumably stability is the key - especially if the complex is only stable in the presence of the additional factor ?
Examples from recent tickets: - receptor/ligand complexes
- protein + DNA/RNA complexes: https://sourceforge.net/p/geneontology/ontology-requests/10319/ https://sourceforge.net/p/geneontology/ontology-requests/10221/ https://sourceforge.net/p/geneontology/ontology-requests/10156/
There is no problem as long as the other molecule is a stable member that is required for the 'function' of the complex. We will try to be coonsistent with InTact for complexes that have more than just proteins.
X metabolism/catabolism to Y
Several requests for these types of terms have come through the termgenie free-form recently. Could we generate a new TG template for these with has_input and has_output ChEBI chemicals?
The templates are ready for testing:
- biosynthesis_from
- biosynthesis_via
- catabolism_to
- catabolism_via