W3C HCLS – Clinical Pharmacogenomics

05 Dec 2012

See also: IRC log




scribenick bobP

<mscottm> Zakim keeps hanging up on me

<matthias_samwald> (probably)

Michel: CSHALS submission, plus...

Matthias: AMIA(?)

<matthias_samwald> IMIA Medinfo

Matthias: Deadlines are coming up. CHALS today, IMIA Dec 10
... we should be at both conferences
... think about integration of resources that we have
... cshals, we can use what we have now; re-use resources from summit next year
... slightly different audiences: AMIA translational, CSHALS more people from pharma

<mscottm> Zakim is telling me that 4257 is not a valid passcode.. ?

Joanne: CSHALS focussed on pharma, but well aware of semweb tech

Matthias: CSHALS relatively short submission

<mscottm> Wow, thanks Zakim for letting me in! with passcode 4257 even.

<mscottm> echo echo

Michel: Happy to do CSHALS presentation
... SWAT4LS was last week in Paris

<mscottm> https://docs.google.com/document/d/1LIO2X7PL3MTHWX5XhIy0Dd8h2yFr0Pk4d8lBG8A6CgY/edit

Michel: Should list resources and dates accessed

<boycer> codehop.dev

Michel: add more than the OWL KB?

Matthias: KB, only referring to the OWL ontology; but includes other data: Michel's data++

Michel: OWL profile?

Matthias: DL

Michel: Can reason, dump inferences to OWLim for query
... or set up website too, or just a file for query

Matthias: Just a file for now
... parts of Bio2RDF + Richard's data, but no real interface

Michel: PharmGKB, dbSNP, OMIM

Scott: What is the biological use case?
... we have snps, structured product labels

Matthais; Show how to integrate different datasets, then demo how OWL reasoning can connect to clinical guidlines

Michel: Brainstorm, w product label for codeine(?)

<michel> that's rich ;)

Michel: statement applied to pregnant women

Rich: Interesting part was that sem model could capture PGx (view)

Matthias: Could use owl for automated reasoning and for matching pts to clinical guidelines
... RDF model for capturing slightly more (subtle rules)
... would be hard to capture everything in clincial rule sets
... OWL inferencing would add here

Rich: +1

Michel: Matthias' point is very good
... RDF can express opaque relationships
... we can cherry pick RDF parts, then construct much more powerful ontological inference

Matthias: 3-4 coherent, not too complex in presentation

<mscottm> http://informatics.mayo.edu/adepedia/index.php/Main_Page

Scott: Another idea: We are bringing in structured product labels; should we address adverse events
... sparql resource for AE reporting systems
... we could use this resource in our scenario

Matthias: but time constrant

Scott: Can use AE pdf w sparql endpoint

<BobF> As full disclosure, the developer of ADEpedia is located directly across the hall from my office :-)

<mscottm> lol

BobF(?): There has not yet been rigorous validation of the quality

scribe: (someone else) addressed validation
... to bring it in, there has to be a connection between PGx and AE

<BobF> (Rich Boyce is speaking)

Rich: Something to think about in the big picture, very nice to address AE outcomes
... 3rd resource is SIDER, slightly different process

Michel: SWAT4LS I wrote a parser for SIDER

Rich: SIDER has 900 drugs, not yet complete
... dbPedia value is the cloud platform
... Jian Yu(?)

<BobF> Qian Zhu - another colleague of mine

<BobF> ADEpedia: Guoqian Jiang

(discussion that BobF is geographically close to ADEpedia :)

Matthias: ADEpedia, RDF does not contain much other than drug -> side effect
... not not much info on (context, provenance)
... still under development(?)

Rich(?): Part of the proposal is to include this provenance

scribe: item for further discussion, to develop metadata standard for structured labelling
... manually annotated AE + structured label: these are claims about the drug
... should be tied to evidence
... to be more concrete, should be SBML

Michel: We had a group on SciDisc, out of which come other ongoing groups
... not too difficult to assemble from these semweb assets

ericP: Anybody else w curated stuff?

Rich: FDA has indexing initiative; there is some validation
... ericP: have done homework on this.

Michel: Trying to link PGx aspects; when talk about drugs we get AE, drug-drug. All connected
... AE, mining structure product label are key endpoints
... for PGx: Do we have what we *can* have?
... PGx has a lot of assertions, needs curation(?)
... excited about pulling out facts w provenance from lots of different sources
... connect to other resources; what is the process to curate the lit?

Matthias: People can go overboard w complex ontologies; 10=headache :)

BobF: One of things to be careful: Separate knowledge from interpretation!
... PGx line greys easily; genotype-phenotype, PGx has habit of tying to intermediate metabolites
... but these are arbitrary, historical in nature but w/o good semantic meaning
... for assertions, consider how much of PGx to include in this project

Joanne: +1 on observations
... worth re-stating this.
... for medical conference

Matthias: Address intermediate metabolite info problems

BobF: Talking re data, resources, queries outside of PGx; but (ex) intermediate metabolite connections between genes may lead us to bad inferences
... primary data sources like pubs, we can have provenance; but then there are lab-to-lab differences
... see this periodically; careful of this trap.

Thanks to Matthias for taking the lead!

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.137 (CVS log)
$Date: 2012-12-05 17:04:33 $

Scribe.perl diagnostic output

[Delete this section before finalizing the 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)

No ScribeNick specified.  Guessing ScribeNick: bobP
Inferring Scribes: bobP

WARNING: No "Topic:" lines found.

WARNING: No "Present: ... " found!
Possibly Present: ADEpedia BobF Bob_Powers GVoice IPcaller Joanne Joanne_Luciano MacTed Matthias Michel P37 P38 Rich Scott Tony aaaa achille_zappa bobP boycer ericP https matthias_samwald mscottm
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Got date from IRC log name: 05 Dec 2012
Guessing minutes URL: http://www.w3.org/2012/12/05-HCLS-minutes.html
People with action items: 

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report

[End of scribe.perl diagnostic output]