Ontology meeting 2012-05-16
- 1 FOLLOW-UPS FROM LAST WEEK:
- 2 DISCUSSION ITEM I: Follow up on "Moving biological_process_xp_cellular_component into live GO"
- 3 DISCUSSION ITEM II (related): Are we going to retrofit logical definitions for cell-type-specific cellular component terms?
- 4 MINUTES
FOLLOW-UPS FROM LAST WEEK:
- Chemical xps
The EBI editors have added all of the "straightforward" missing cross-products. The only ones left are those that require a new CHEBI ID, and the ones that we haven't resolved yet on how to handle. For the former, the EBI editors will submit the needed new CHEBI terms using the direct ChEBI submission tool. For the latter (e.g., how do we handle "via" terms?), Jane is going to write all pending questions/issues down in a wiki, and we can update it as we go along.
Q: How are we going to handle periodic reviews of CHEBI terms? E.g. if we submit a CHEBI term, then use the CHEBI ID in xps, but then ChEBI reviewers don't promote the CHEBI entry?
Q: Are we going to do a global assertion of new links inferred by CHEBI? And a removal of links that cannot be inferred? How to co-ordinate this?
- TermGenie and ChEBI
TG now writes to both the main GO and to the xp file. The template has been tested, synonyms and Dbxref field have been added.
Curators can now use the template themselves for M/C/B when the chemical is in ChEBI.
- Next templates for TG:
Binding, transport, response to, X biosynthetic process via Y - these all have fairly complete sets of xps already so should be straightforward?
What next - substrate-specific transporters? Response to chemical substance?
- ChEBI paper
- TermGenie paper
(On a related note, we might want to plan implementation of new types of TG templates. One that would be of frequent utility is "response to X/cellular response to X".
E.g. see this recent SF request:
DISCUSSION ITEM I: Follow up on "Moving biological_process_xp_cellular_component into live GO"
See previous discussion here: http://wiki.geneontology.org/index.php?title=Ontology_meeting_2012-03-07
- I'm ready for all the transport ones - one hold up has been thinking of a family of relation names that generalizes over both transport and localization that encompasses the different roles (cargo, source, destination, conduit). If we decide on these we can just move this lot in straight away. There was a thread on the list about this w/ Midori a month or two ago -- cjm
Some of these (e.g. neuron projection) are in the scratch directory as cellular_component_xp_cell.obo.
Stems from this SF item: https://sourceforge.net/tracker/?func=detail&aid=3518879&group_id=36855&atid=440764
Also see this other SF item: https://sourceforge.net/tracker/?func=detail&aid=3410166&group_id=36855&atid=440764
- Cam-Eds have been adding in missing XPs (hundreds) for:
- biosynthesis/catabolism from x/to y
NEXT SET OF CHEMICAL XPS TO ADD
1/ BIOSYNTHESIS/CATABOLISM VIA
- use 'has_intermediate'
- Jane to divide the approx 90 terms between us.
2/ TRANSPORTERS (MF)
- InterPro are about to start requesting lots of substrate-specific transporters for their mappings, so we should get these in soon.
- Most are 'transmembrane transporter' terms; NB: some of the children don't have TM in the name, but they do have a 'TM transporter' parent. For these, we should add 'TM' in as exact synonyms.
- For TM transport/TM transporter terms, we all agreed we should use 'transmembrane transport' as the genus, to keep it simple.
- scratch/xps/molecular_function_xp_chebi.obo has some transporter activity XPs. They need a review given that they were done automatically and a while ago. Chris will automatically generate these again.
AI: CHRIS TO AUTOMATICALLY GENERATE NEW TRANSPORTER_CHEBI XPS
- Q: Can Heiko go ahead and add the templates in for chemical binding and transport?
- BINDING: has_input, so YES
- TRANSPORT: we need to decide on relation names, and make them generic enough that we can use them for localization terms too. At the moment, we have 'results_in_transport_of' but we need something more generic than this because not all establishment of localization involves transport.
3/ RESPONSE TO (our favourite!)
- These cover both chemicals and stimuli (e.g. stress). We're still not keen on these broad terms, so will come to these XPs after we've done transporters.
PLAN OF ACTION FOR ALIGNING GO WITH CHEBI
- Cam-Eds to finish adding in XPs
- Assert all the links suggested by the reasoner report.
- For the WRONG links suggested by the reasoner report, it's trickier because some follow the role hierarchy. Therefore we need to go through and tease out the role ones that we want to keep, and the true incorrect links that should be binned.
- Q: How are we going to handle periodic reviews of CHEBI terms? E.g. if we submit a CHEBI term, then use the CHEBI ID in xps, but then ChEBI reviewers don't promote the CHEBI entry?
- A: An ALt_ID in the XP will render the XP 'silent' so that's the worst that will happen. We could introduce a step to re-run the script every time ChEBI does a release.
- AI: David and Tanya to force Chris to finish it when they visit Berkeley on Monday.
- AI: Heiko should be added to authors. Then Heiko can look-over the draft from Chris next week.
- AI: Need to send final draft to Mike Bada.
Nothing exciting to report.
BP_CELL COMPONENT into live GO
- Needs more discussion time as we need to think about the relations.
- David, Tanya and Chris will work on this on Monday.