HCLSIG/PharmaOntology/Meetings/2009-03-19 Conference Call

Conference Details

Date of Call: Thursday March 19, 2009
Time of Call: 12:00pm Eastern Daylight Time (EDT), 16:00 Greenwich Mean Time (GMT)
Central European Time (CET)
Dial-In #: +1.617.761.6200 (Cambridge, MA)
Dial-In #: +33.4.89.06.34.99 (Nice, France)
Dial-In #: +44.117.370.6152 (Bristol, UK)
Participant Access Code: 42572 ("HCLS2").
IRC Channel: irc.w3.org port 6665 channel #HCLS2 (see W3C IRC page for details, or see Web IRC)
Duration: ~1h
Convener: Susie

Agenda

Review Actions
Finalize categorization of questions - Christi
Lilly's PR&D Ontology - Jenna Zhou
AMIA Trip Report - Susie
F2F plans - Susie

Attendees: Larisa, Jenna, Lynn, Christi, Elgar, Susie, Matthias, David, Colin, Codi, Trish

matthias_samwald: Susie: at the moment we are creating a high-level template (ontology)
matthias_samwald: ... it does not have workflow component at the moment.
matthias_samwald: ... but this might be added when we apply it on the use cases
matthias_samwald: Christ: reviewing the areas....
matthias_samwald: ... project manager role was added
matthias_samwald: ... has only a few questions. maybe it should be removed?
matthias_samwald: ... within-company focus vs. cross-industry focus
matthias_samwald: ... for some others (e.g. biology, physiologists)... maybe they need to be collapsed? are they not generic enough, e.g., specific to a certain company?
matthias_samwald: Colin: do we have contacts/representatives for each of these roles that we can ask, inside the pharma ontology team or at least in its proximity?
matthias_samwald: Christi: we need feedback on whether is also valid for other companies.
matthias_samwald: Susie: many of the questions have originiated from Lilly, and from Lilly's perspective the questions seem to be on target.
matthias_samwald: Susie: but we need feedback from other companies.
matthias_samwald: Larisa: if find it very difficult to judge, some terms (e.g. "outcome) have many meanings. the terms should be clarified.
matthias_samwald: Christi: maybe we can add term definitions to the wiki
matthias_samwald: Susie: a lot of that would come out when we identified the top 30-40 terms and start thinking about how they should be related
matthias_samwald: ... then we need a crisp definition.
<matthias_samwald> ... i guess it would make sense if the questions can be as self-describing as possible
<matthias_samwald> ... i am weary of spending too much time on this at this stage, because this might result in problems with our timeline
<matthias_samwald> Lynn: some of the roles are a bit overlapping.
<matthias_samwald> ... a general care physician might refer a patient to a specialst... but the types of questions they are asking would be the same.
<matthias_samwald> Christi: do we have consensus to collapse those roles?
<matthias_samwald> Susie: I think the specialst would be able to ask more detailed questions. at the moment this is not clear in our questions, because we have not chosen a specific therapeutic focus.
<matthias_samwald> ... even if two sections are identical, it helps us view things from different perspectives. this way, we are less likely to miss something.
<matthias_samwald> ... example for a GP-only question: "what specialst should i refer this patient to?"
<matthias_samwald> lynn: specialsts ask specific questions AND basic questions, whereas GPs only ask basic questions
<matthias_samwald> Susie: I think it makes sense to have a project manager role
<matthias_samwald> lynn: the marketing role would also be important
<matthias_samwald> colin: i have been trying to add these
<matthias_samwald> susie: trish, could you review ontologies?
<matthias_samwald> trish: bioportal has >150 ontologies (hosted by NCBO)
<matthias_samwald> ... you can use the search function to identify fitting ontologies.
<matthias_samwald> ... you can also do it in batch
<matthias_samwald> TODO: trish will review ontologies that would fit (over the next weeks)
<matthias_samwald> TODO: colin will help with reviewing ontologies, especially chemistry.
<matthias_samwald> Christi: i just tried bioportal, it's great
<matthias_samwald> trish: there is another tool where you could submit text that recommends fitting ontologies
<matthias_samwald> trish: "Ontology recommender service"
<matthias_samwald> trish: http://obs.bioontology.org/oba/Recommender.html
<matthias_samwald> Jenna (lilly): i have been working on an ontology
<matthias_samwald> ... the project started last year
<matthias_samwald> ... we tried to connect the data silos
<matthias_samwald> ... four major entities/classes: project, assay, molecule, target
<matthias_samwald> ... any questions?
<matthias_samwald> ... i reviewed papers. most people start with questions. in our project we did not have these questions, but data.
<matthias_samwald> ... we did entity-relationship modelling
<matthias_samwald> ... scientists verified these relationships
<matthias_samwald> ... most of these things are straight-forward
<matthias_samwald> .... "target" is ambiguous, though.
<matthias_samwald> colin: it sounds more like a goal of research
<matthias_samwald> ... calling it a target might confuse people
<matthias_samwald> jenna: i mostly modelled everything with subclass relations
<matthias_samwald> susie: can you provide examples of questions that you can now answer?
<matthias_samwald> jenna: "what targets have been tested for this project".
<matthias_samwald> ... targets are linked to assays
<matthias_samwald> ... which are linked to molecules
<matthias_samwald> ... the ontology is growing organically, new things are added upon customer requests
<Susie> Colin: Have you loaded the ontology into Protege
<Susie> Larisa: Works in Protege 3.3
<Susie> Larisa: Many relations in the ontology, but they are different from OBO relations, so will be hard to integrate going forwards
<Susie> Colin: In ontology, things done as properties, would best be done as roles
<Susie> Colin: Lots of work needs to be done on properties
<Susie> Colin: Not that it's bad...
<Susie> Trish: Are you familiar with RO
<Susie> Trish: Easier interoperability if use RO for relationships
<Susie> Colin: Useful start look at the BFO
<Susie> Colin: Helps separate our roles, processes, etc.
<Susie> Larisa: I'd be happy to help with the ontology
<Susie> Jenna's email address for feedback is jennaz@lilly.com
<Susie> Christi: What would you call a 'target'?
<Susie> Larisa: A 'target' would be an 'objective'
<Susie> Larisa: Can have 'Subject of Target'
<Susie> Larisa: Andrew Hopkins is working on an ontology of targets
<Susie> Larisa: The Lilly ontology looks great, but could be re-arranged a little to be a bit more compliant
<Susie> The AMIA Translational Bioinformatics Conference was earlier this week.
<Susie> Was attended by approx. 300 folks from a diverse range of backgrounds
<Susie> There was a well attended session on Advanced Web Technologies that included Kei Cheung, Matthew Scotch, Tim Clark, Nigam Shah and myself
<Susie> Much interest in Linked Data and other topics
<Susie> The HCLS F2F meeting is being held in Cambridge on April 30, and May 1
<Susie> There are invited talks and task breakout sessions
<Susie> I'd like there to be a Pharma Ontology breakout session
<Susie> We could decide on the 3 use cases that we want to write up, as well as start identifying key entities
<Susie> The agenda and registration are on the wiki
<Susie> Susie: F2F http://esw.w3.org/topic/HCLSIG/Meetings/2009-04-30_F2F