XP:Meeting XPs 2008-12-11
2008-12-11 8.15 am PST
DavidH, DavidOS, Jane, Midori, Jen, Amina, Chris
Hi Chris - Jen, Midori and I had a meeting about implementation cross-products this morning. We've come up with a provisional plan to augment what you sent out which we've posted on the wiki: Editors_cross-product_implementation_plan We've also got a few questions/issues we were hoping you could address for us:
1. We don't think it is possible using current OE2 filters to save out just the xps, so for the transition phase we'll probably just edit the xps OBO file by hand and commit to cvs.
For this reason, and so the xp definition files don't get too out of date with the live ontology, we want to get the xps in the editors file pretty soon - we're thinking the beginning of Jan. Could you and Mike C work out a mechanism to strip the intersection tags from the editors' file during the file-release process?
Also, we think we'll eventually have users' versions of the GO .obo file with and without the intersection tags. Is that correct?
2. OBO-Edit/implied links: if we want to use OE2 to use the xp defs to make new links, we run the reasoner (which one? The new link pile reasoner?) and then save the file with the links - to do this do we check the 'save implied links' and/or the 'realize implied links' boxes when we save? Should the reasoner be on during the save? Do we need to use any other filters when saving this file?
3. The imports file link is broken in this page:
4. How do you want to handle promotion of the xp definition files from scratch into cvs 'proper'? Do you want us to copy them into a new directory: go/ontology/cross_products once they've been checked?
David and Tanya - copied you in as a heads up as we've 'volunteered' you to maintain certain xp files - see the wiki page. We'll send an email around the rest of the ontology editors very soon.
Jane, Jen & Midori
Some ideas for metrics:
- coverage (what % of the terms are covered)
- genus terms: are they consistent? Correct?
- differentia filler terms: do they point to the correct term
- differentia relations: are they consistent? Declared? Defined?
- Definitional correctness
- Consistent with text definition?
- necessary and sufficient?
Relation Progress (Chris)
- Many relations used are not declared (eg in a stanza in ro_proposed), causing OE problems.
- Even if declared, relations may not be defined, or defined well.
- ro_proposed is a bit big. Would be better to load in only those actually required
- Added *.relstats file as a stopgap. E.g. biological_process_xp_cellular_component.relstats
Inidividual XP sets Progress
Currently being tagged with ! comments in the obo file (which would be lost in an OE roundtrip)
- Thread: CC organization question for xps
- Meaning of 'anatomical structure'
- XP:cellular_component_xp_go (Midori)
- surrounds, surrounded_by
- bounds, bounded_by
cytoplasmic X does not have part cytoplasm!
Progress notes: Biological_process_xp_self_ProgressNotesNov2008
Too big? Split out somehow? Currently causing memory problems for even quite fast computers.
- Various regulation xps (DavidH)
- fly_anatomy_xp (DavidOS)
OE Tips and Issues
- Rendering XPs - should we have a better way to share configs?
- Reasoner update - Chris. See OBO-Edit:Reasoner_Benchmarks
- Which reasoner should you use?
- Memory problems? Note which machine your using.
- Are we happy with the no-edits-in-OE constraint for the time being?
- Timetable for integration into GO?
- How will end-users see these?