<ericP> scribenick: bobP

Michel: Roundup of the task forces
... Linked Life Data. working on two W3C notes
... 2nd note is Expression RDF; sparql to query multiple endpoints w data in different forms
... trying to write W3C note for conceptual framework for expression data, standard representation
... working w Gene Expression Atlas
... have a nice diagram. Next meeting is Tues
... Groups that have pathways etc, we would be drafting notes
... Next is Helena to link life data, figure out transformations to complexes, relationships to diseases etc
... work is in progress, busy at DERI.
... I will visit DERI at end of Nov. Another group has released new sparql editor
... also working on a cloud platform
... Systems Biology. Working closely w Matthias in clincial decision support
... trying to overcome the problem of alleles and genotypes
... people categorized as *1 or *2; but as more SNPs discovered, allele definitions are renewed
... Bob Freimuth has come up w OWL reasoning solution, even if the nomenclature changes in the future
... working on an AMIA presentation
... Mattias also has Medical Microdata Compendium; should explore w linked open data sets
... using sparql and PGx
... This is a quick run-through.
... Want to talk about Charlie and ericP are working on.

<ericP> TF page

ericP: Current COI task force page

:)

reviving COI

scribe: Goal of COI was to address: given a lot of standards, and need interop, then if standards were based on RDF interop would be easier
... notion was to come up w a demonstrator
... I created what HL7 would look like in RDF, (straw man who stood up :)
... demonstrator did have the cool effect that, given RDF, here's how we can get interop
... to take a clinical trial, express as query in language familiar like CDISC-y
... CRO people can write these kinds of queries
... But language of hospitals is HL7
... n clinics in HL7-like; write a Xform from pattern -> pattern, pushes query into CDISC-like
... Xform through three different materializations to do this pipeline
... shorter: use queries within sparql construct rules (?)
... This worked over one corpus of clincial records; success w inclusion and exclusion criteria
... queries can get large, but they are quite effect b/c they go directly to the RDB
... This was the initial project at COI. Perform recruitment queries for trials
... Working now w Lloyd McKenzie(?) to turn HL7 into OWL
... he is xforming UML into OWL(?)
... we get for free an RDF representation of RIM
... used also to define the wire format FHIR
... data can be either true RDF or grrdl-able into RDF

<egombocz> apologies to all - need to leave for another meeting

scribe: Another feature is Charlie's use case: snomed here, hl7 there, still a big space in how clinicians write things down
... need to establish equivalences through longitudinal records

<michel> http://www.amia.org/jointsummits2013/proposals-types

<michel> podium presentation due oct 11

Q: There is a goup in HL7 Clinical Interop Consortium

Charlie(?): One of the goals is to decide RIM/ SNOMED/ etc

scribe: would like to get CIC and COI talking together; working on the same problems

Kersten: FDA has an interest in next-gen exchange formats
... talk about HL7 v3 format and others
... FDA asking for inputs on other options; I am following along on a series of interesting discussions
... ericP should be in their coming meeting, like w the FHIR work

<kerfors> Turning CDISC standards into RDF/OWL & SKOS and also using ISO11179. Joint presentation AstraZeneca and Roche at the CDISC Europe conference earlier this year http://kerfors.blogspot.se/2012/05/semantic-models-for-cdisc-based.html

Kersten: Also I work in the CDISC community
... Roche has all CDISC and internal, all in RDF
... time to review that work. Would be good if we had all different standards in RDF
... CDISC conference late Oct Baltimore
... we have put together an open discussion w Roche and we are going to do what we say
... also looking for way to publish this in the short term
... then push RDF work back to CDISC
... News that there is a new pre-competitive org Transcelerate
... pharma thinks cdisc too slow, maybe Transcelerate can push faster
... Roche work meeting next week

Scott: Interesting that Transcelerate has emerged; to push important work

Kersten: Frank H. will review

ericP: Keep our eye on: Work w Roche CDISC, and they have their own defs of observations etc
... those are based on HL7 v3 observations
... we should make sure that defs are dropped-in pieces

<kerfors> Also, an interesting presentation from CDISC SHARE and GSK on the problems with today's data standards and the idea of using "mind maps" for the next generation of "concept based" standards http://kerfors.blogspot.se/2012/09/mind-maps-just-begging-for-rdf-triples.html

Charlie: Then we are sitting right in the middle of the anaphylaxis AE problem.

ericP: Work for normalizing forms will be relevant to both trials and care

Charlie: IMI(?) has been building solution?

Kersten: Pushing, but it's not easy.

<Lena> ...ehr4cr is more hl7/rim type of model as opposed to semantic web oriented...

ericP: Anaphylaxis use case, any successes will have immediate applicability
... with the RDF representation, we can turn the crank; CRO observations/hospitial observations

Charlie(?): BRIDG is subset of RIM

ericP: Lloyd mailed an OWL representation of BRIDG

Charlie: Lloyd will re-do the BRIDG-RIM map in OWL

ericP: Cool.

Michel: Will do two AMIA submissions
... This meeting is adjourned.

rrsagen, draft minutes

This is scribe.perl Revision: 1.137  of Date: 2012/09/20 20:19:01  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found ScribeNick: bobP
Inferring Scribes: bobP

WARNING: No "Present: ... " found!
Possibly Present: Bob_Powers Charlie IPcaller Kersten Kerstin_Forsberg Lena MacTed Michel Scott aaaa aabb aacc bobP egombocz ericP erich kerfors scribenick
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Got date from IRC log name: 02 Oct 2012
Guessing minutes URL: http://www.w3.org/2012/10/02-hcls-minutes.html
People with action items: