Ontology meeting 2012-08-08
ATTENDEES: Paola, Becky, Jane, Harold, Tanya, Chris, Heiko, Judi
UPDATE ON CHEBI PAPER
(Carried over from last week)
We had initially left these aside, but we have more and more requests coming through SF, so we need to work out a strategy for adding these in via TG.
(There are some xps in an old file called biological_process_xp_stimulus.obo in the scratch directory)
Chris wanted to merge the xp-files into the main ontology file before doing this. We should work out the genus and relationships that we want to use in the logical definitions. We should also use protege to check to see if the existing terms are placed properly. (David & Tanya please report on this). Currently there is nothing to report because David is on holiday.
For the response to chemicals, we could work bypassing an external stimulus ontology.
- A related issue: parentage to "response to drug" vs. "response to chemical stimulus".
At the moment, we're not being consistent on how we do these. Related SF items:
We'd have to refer to roles as well. We'll make a couple or more of new relationships for it.
AI: Chris will do a first pass to pull out/parse the chemical stimuli terms, from the GO file itself. (We'll forget about stimulus.obo for now.)
FOLLOW-UP: EC numbers in GO - feasibility study
See Action Items and background here:
NOTE: On July 12th, Rhea made a new release (31):
"A new Rhea release is available at http://www.ebi.ac.uk/rhea with 19471 unique reaction IDs.
NEWS concerning downloads: fields added to RD files; new BioPAX 'lite' file; cross references exported as TSV files."
The Rhea downloads page is at: http://www.ebi.ac.uk/rhea/download.xhtml?
AI: Chris and Heiko to follow up. Revisit in one or two weeks' time.
NOTE: New Rhea release (32) on Aug. 9th, see http://www.ebi.ac.uk/rhea/ (Paola)
ADDING NEURON TYPES IN THE CELL ONTOLOGY
The NIF maintains a fairly comprehensive list of neurons (135 of them), arranged by type of neurotransmitter:
Many of these cell types appear to be missing from the Cell Ontology. However, the "parent" neuron types from NIF are already in CL (except for one or two that can be easily added).
Shall we consider adding to CL the missing neuron types from NIF? We could place them simply as is_a [neurotransmitter-type] neuron (e.g. GABAergic neuron), and part_of [anatomy location as UBERON ID]. The is_a parents can be taken care of automatically (the NIF list, complete with definitions, is downloadable in CSV format). The part_of parents should also be fairly straightforward and requiring only a minimum of manual work. We could use Populous. Could this be a job for OPPL alternatively?
AIs: For the time being, only add the CL terms that are needed for corresponding GO term requests and cardiac conduction work.
Judy will keep us posted about a larger effort.
Paola will add CC terms from SAO in September.
CELLULAR PROCESS CAUSING PROBLEMS AGAIN
- The pesky 'cellular process' term is causing problems for several of the virus terms. Particularly 'DNA metabolism' and the terms around it.
- E.g. 'DNA integration ; GO:0015074' often involves integrating DNA from one organism (e.g. a phage or other virus) into the genome of another organism. With 'DNA metabolism' current a child of 'cellular process', the child term 'provirus integration; GO:0019047' can not be given a 'multi-organism' process parent.
- Options include:
- Moving 'DNA metabolism' out from under 'cellular process' and redefining it (and any children) so they're not limited to cellular processes.
- Creating cellular and generic DNA metabolism terms... could get very confusing and redundant.
- Taking a fresh look at what we mean by 'cellular process' and how this fits with the multi-organism processes.
- Is this really a problem with cellular process or is it a problem with a disjoint relationship between cellular process and multi-organismal process? When we originally thought about this, we were considering things like mating etc., not viruses. I actually think at the level of viral processes, a process can be both multi-organismal and cellular. (David)
Chris: another option is saying that viruses aren't organisms
Create an additional top-level split between e.g. 'single organism process' and 'multi-organism process'? We need a feasibility study.
AI: Jane and Becky will look further into this, and report back.
AI for a next meeting: discuss BFO.