<mscottm2> beautiful music? supplied automatically by hold-function of phone I suppose

<iker> yes, nice music

<matthias_samwald> scott bauer probably

<matthias_samwald> yes!

<adrien> I was talking

scribenick bobP

Matthias: TMO for Pgx use cases

<michel> http://www.w3.org/wiki/HCLSIG/Use_case/TMO_for_pharmacogenomics

Matthias: Decided should be organized around small manageable use cases
... up to 12 months (at most?)
... overarching goal of using semweb to enable personalized medicine
... one is extending TMO to cover stratified Pgx
... to continue this vertically integrated TMO
... should try to agree on extensions by EO Feb
... should try to map datasets to this new version
... ICBO deadline Jan 29

<matthias_samwald> http://code.google.com/p/translationalmedicineontology/

Matthias: main obstacle now, need to find some mode of working together, co-ordinate on what's in ont
... good objective, to collect models of Pgx, look as SNP ontology, etc
... make comprehensive list of such models

<ericP> Life Science Domain Analysis Model

ericP: Extends TMO: do we have queries in mind, to help get a sense of data?

Matthias: Started out by creating a table of competencies that the ont can handle

<iker> if you could paste it to the wiki that would be great\

ericP: We gather info, but sometimes not actionable(?)

Scott: Refer back to original review article; essentially motivated by CDS
... where clinician has SNP info, like Harvard videos
... but most of web is buried in deep web; build on sparql instead of screen-scraping
... use case based on situation not real today, and also requires regulatory input

ericP: Before, we had a hard time talking about input to physicians

Scott: In clinic, personalized med hasn't reached thru, but it's on the way.

<Zakim> ericP, you wanted to ask about concrete motivating queries

Scott: added value is subtle. sparql can more easily link to data

Matthias: True, not very established at the moment
... but look at the future. Vienna is trying with CDS, collaborating w group in Estonia
... Estonia capturing data in a biobank, available in a wide open system
... maybe we can realize such a system (someday? :)
... another scenario, adoption of genetic testing for cancer treatment

<matthias_samwald> cancercommons.org

(this needs tissue-specific expression data!)

Matthias: Melanoma, colorectal thru semantic mediawiki

<ericP> i think matthias_samwald's point about "taking the extra step" for grave diseases bears relevence to where we can demonstrate and advantage to having a semantic web of clinical data

Lena: +1 to Matthias. New generation of physicians are completely aligned w pers med
... also, not enough to look just at SNPs. Translocations, exomes, non-coding regions, all of this important
... if TMO focusses just on snp, will miss the point.
... emphasizing that SNPs are not enough!

<iker> i agree with Helena

Scott: Yes, we are just starting w SNPs, but will continue w modelling

<iker> ok

Lena: SNPs low hanging fruit has already been picked!
... Ex DNA repair mechanisms. (talking complex pathways in cancer cells)
... *not* just about broad genetics. It'w what goes on inside the cell

ericP: Need to demonstrate that we're not limited to SNPs
... Small no. of things?

Matthias: TMO should not head this way
... we need to show path to CDS
... from perspective of clinical application, seen as plain biomarkers

Lena: So that's a biomarkers ontology, not a translational medicine ontology.

(Lena holding forth here!)

ericP: What to model and what not?
... if goal is to enable CDS, then model kinds of data that we can write rules about w today's literature

<mscottm2> I think that Mayo Clinic participants can speak more directly to what is most relevant to CDS

ericP: clinical use case different from research use case(?)
... low benchmark: clinical practice to model cutting edge questions of today
... higher research: we need more info in pool
... people doing specialized research won't see value in integrating data

<mscottm2> 1+ ericP

ericP: but clinical integration story is more helpful

Lena: This is a good strategy.

ericP: Arguing that b/c we integrated data, we were able to ask these questions, answerable in two years
... to maximize our impact
... to get a known audience

Matthias: Mayo Clinic can help us to see what's interesting in a 7-year time frame

BobF: +1 to ericP, defn should be useful 2-3 years from now
... Mayo couple of pilot project to look at Pgx-to-clinical, but they are still pilots
... SNP is first we look at at Mayo
... after SNP, then we will expand to other types of genomics data

ericP: Pilots: Integration, actionable, etc. Can we glom onto one of these :)

BobF: Thinking. Need to consider further.

Scott: Synergy w Linked Clinical Data LCD project

BobF: LCD is exploratory, primary focus has not yet been clinical
... but he is thinking about ways to bring in genomic info

Michel: Lena, you want to capture all these other indicators
... also, some of them are actually being used for CDS

<iker> I must leave too, see you folks

<mscottm2> bye

Lena: Can't limit ourselves.

Michel: Looking at time frame, but should consider these other data types too

Lena: Some mechanisms can silence SNPs.

Michel: Approved procedures may not include these
... ultimately we do want all of the data to come into the analysis
... should look at what the next datatypes are
... and how they have been linked to clinical outcomes

Matthias: Channel this energy! to develop TMO etc

Lena: Why not go after FDA-approved SNPs only?

Matthias: FDA-approved SNPs in package inserts. Maybe we should limit ourselves to this.

<michel> +1 for FDA approved SNPs

Matthias: this has warfarin plus many others

<michel> including warfarin.

Matthias: good idea
... Keep working on google doc
... should also finalize TMO extensions asap

Michel: What is total list of terms? What's in, what's out
... so we know when we're done.

Matthias: Tried net Protege w Iker, but the SVN seems to be more popular

Michel: Consider all the terms for addition

BobF: Driven by use cases?

Michel: Still need a list. Started TMO w speadsheet of terms

Matthias: This way, people start to add topics, not nec translatable to classes

Michel: Use cases -> topic lists, need to have the list
... check off terms as they go into TMO

Scott: We were guessing that SNP info, are there interesting things to check out
... warfarin: practical to be able to look up the dose
... in clinic w regulated knowledge. Ethnicity can be a partial substitute for SNP info
... very simple scenarios; drug, patient type, diseases, therapies
... loaded copy of dbSNP, trouble getting disease related from SNP
... correct in this??

Lena: Looked at chain for cyp2c9. Look at clincial significance: unkown, not available!
... physicians must know where the data are

BobF: Most of that info is *not* put into dbSNP
... there is another site w dosing info
... keep in mind, as at Mayo, there is clear distinction between info clinical guidlines and guidlines themselves(?)
... this kind of integration of TMO will be more important for guidline development than for clincial practice(?)

(getting this right??)

scribe: guidline development will use this more than clincial practice

Lena: Maybe physicians can push FDA

BobF: CPIC is working on developing such guidelines
... agree that this would be a good place to start. CPIC would give confidence that we are covering the space
... how applied: How data are generated to begin with

<michel> sorry i got kicked

<michel> and can't rejoin

Scott: We had created two categories, clinical practice and research

Matthias: So we focus on FDA-approved SNPs to start
... BobF also mentioned that specific guidelines might be useful in developing TMO
... also his good idea, use case for physicians on committees to develop guidelines

Scott: Microarray profiles are used in guidelines
... profiles used now for which chemo to use

Matthias: Create another google speadsheet for additional terms for TMO
... 1) label 2) description of term; Maybe copy existing TMO to bottom

<matthias_samwald> https://docs.google.com/document/d/1I9xyVKhO9wG7My2fWu9S-wMwWZxG6PkRc98kOxUqH-o/edit?hl=en_US

Matthias: Future Medicine response to reviewers

<matthias_samwald> https://docs.google.com/document/d/1Vn-9nzlGJXoO1NE3pWGMrRFpigB2qaji7D-27lNcGoc/edit?hl=en_US

This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
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: BobF Bob_Powers IPcaller Lena Matthias Michel Scott Scott_Bauer aa aabb aacc adrien bobP ericP hcls2 https iker joined matthias_samwald mscottm2 rickkiefer
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting


WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 01 Dec 2011
Guessing minutes URL: http://www.w3.org/2011/12/01-hcls2-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