HCLSIG/PharmaOntology/Meetings/2009-10-12 Conference Call

From W3C Wiki

Conference Details

  • Date of Call: Monday, October 12 2009
  • Time of Call: 10:00am - 11:30am EDT
  • Duration: 1.5h
  • Convener: Elgar, Joanne

Agenda

  • Technical issues of TMO management
  • What needs to be done next and who does it?
  • TMO content

Minutes

Attendees: Colin Batchelor (C), Michel Dumontier (M), Joanne Luciano (J), Elgar Pichler (E)

[Action items are marked by ==> and are included in the minutes at the time they were discussed.]

Technical issues of TMO management

E: How was tmo.owl created?

C: So far built in Emacs, directly editing RDF/XML. Want to migrate to Protege4. Have done initial load into Protege4.

E: What about informative display? So far unable to do label instead of name display in Protege3 - the same works in Protege4 though. Anybody have an idea how to do this?

J: No working experience for this. In general, P4 better for working with OWL, inferencing better in P3.

E: Need to ask how to fix display problem. Also what to do for diffing of versions to see incremental changes. Know how to do this via Prompt plug-in in P3. P3 also has collaborative capabilities, so edits could be nicely commented and reviewed. How about using that?

J: Can ask Matthew Horridge how to do diffing in P4.

==> J: Ask MH about P4 project diffing.

E: Michel also seems to know how to do this. Ask him.

==> all: check with Michel how he did/does diffing.

C: Done so far: input curation status markers as in OBI; a few more definitions are needed; move over to IAO standards; classes left to be done are mostly at end of OWL file (e.g., biomarker).

E: What is status of classes?

C: curation status can be: metadata (in)complete, pending for final vetting, ready for release, ...

J: talked about various status choices in one of last HCLS TMO weekly meetings

E: will look them up; assume no classes are ready for release; what is deadline for getting them there?

C: none are ready for release; deadline should be next Friday because of paper submission deadline

What needs to be done next and who does it?

E: What are priorities for work; who should do what to avoid overlaps and conflicts and to maximize use of resources?

C: Need to agree on data sets. That will determine meaning/use of some classes. Have tried to impress on Susie that we need agreement on data sources

[Michel joins TC.]

M: Before we do anything else need to create external.owl file for ontologies imported by tmo.owl.

E: tmo-external.owl contains frozen content of all ontologies directly and indirectly imported by tmo.owl. tmo-external.owl will be distributed with tmo.owl. Having a local copy of imported external ontologies will making working with tmo.owl from behind firewalls easier and will provide stable external ontologies during development.

==> [priority 1] M: create tmo-external.owl and adapt tmo.owl accordingly

E: After that can finally add class xrefs.

==> [priority 2] E: add class xrefs to tmo.owl

C: can then add further xrefs from Joanne et al. to tmo.owl; maybe also add relations

==> [priority 3] C: add further xrefs to tmo.owl

==> [priority 4] C: add relations to tmo.owl

C: concerned about use case & data sets

TMO content

E: What has been done so far?

C: Took ~50 terms that were collected in spreadsheet. Added definitions when available. [E: Defined isA relations, integration with BFO, use of IAO, ...]. Note that some/all terms might devolve to OCR terms when that becomes available. Maybe state in paper that it is a goal of this work that terms in TMO become subsumed by other spe-cial-ist ontologies as they become available.

J: some terms exist already in MeSH, SNOMED, ...

C: Problem with MeSH; not ready for ontology. Might want to point out in paper that we support "ontological coercion" [please explain/define term again]; term should go into paper.

J: support that

E: What needs to be done, how? Noticed classes under Thing, where should they go? Should we move classes that seem "misclassified", how to do other annotations? What is process of coming to agreement on changes? Who makes next/final merge.

C: We need data set agreement to properly place/define classes directly under Thing.

E: Shouldn't ontology be independent of the data set. Should be able to use TMO for different data sets.

C: E.g., Outcome could be a Quality or Disposition, need example for that. Other cases like that.

all: discussion on change strategy

==> all: next steps for ontology changes - execute in this order:

  • Michel cuts new version of tmo.owl; provides also tmo-external.owl; uploads both to Google Code SVN repository
  • all check out new version
  • all review ontology and make comments in local sandbox copy of owl.tmo
  • all local sandbox copies with new comments are sent to Colin by Wednesday (2009/10/14)
  • Colin attempts merge of sandbox copies
  • discussion of comments occurs on Thursday (2009/10/15) at the earliest (at weekly TMO meeting?)
  • after agreement is reached changes are made to tmo.owl and new version is uploaded to Google Code SVN repository
Retrieved from "https://www.w3.org/wiki/index.php?title=HCLSIG/PharmaOntology/Meetings/2009-10-12_Conference_Call&oldid=10151"