scribenick BobP

<iker> zakim. P50 it is me

<matthias_samwald1> http://www.w3.org/wiki/HCLSIG/PharmaOntology/Meetings/2011-06-09_Conference_Call

Matthias: Agenda: text mining, from spreadsheet to RDF
... Iker has posted RDF

Not really useful to xform spreadsheet to RDF

Iker: Have put up new RDF
... did not understand spreadsheet quite, but did try to xform
... consequences from pubmed articles?
... there are some fields w numerals, other fields w/o meaning

(there's a terrible echo)

scribe: did not find all info in the spreadsheet

Matthias: The tool for spreadsheet in google docs also capable of xforming to RDF

Two outputs, so can see both

Matthias: Iker, you are using Google refine, may want to use G refine again in future
... Agenda: Ontology. Wiki has two arenas for ont development
... webProtege, and dedicated Google doc where Michel has collected ideas for axioms, etc
... Michel and I have been collaborated, but on the separate platforms
... work-in-progress. Tried to describe genes, SNPs, etc under TMO.
... not as abstract entities, but as material entities w. partonomy, haplotypes, nucleotypes etc
... accounted for polymorphism across organisms; genetics compatible w TMO
... difficult to do this. Under heavy development. Question whether we can really do it this way.
... maybe represent all of this as some kind of information artifact rather than material.

Adrien: Will look. How do you re-use SeqOnt?

Matthias: Seems like we are still experimenting to forumlate as TMO-ish

<iker> my regrets but i must leave, I apologize

Matthias: SeqOnt talks about sequences, in the abstract. Matierially it introduces interesting problems.

<matthias_samwald1> https://docs.google.com/document/d/1BeaU3VJW1S_R3eB5ncrcW9wJ-lMU4xdmxCdjH8E0i5s/edit?hl=en_US

Matthias: difference bet gene and allele. Represent all as material then difference between gene and allele starts to disappear
... You can see here some of the things that Michel came up with.

<Joanne> 10 minute warning. Am at EBI and leaving in 10 minutes.

Matthias: if all material entities, then notion of gene starts to disappear

Hi Joanne!

Matthias: This gives the notion of what some of the issues are.

<Joanne> my update: I looked at my notes of what I converted and they are too low level to present. Tim (used his tool) has not been available and I haven't had time to run through his documentation (I needed the brief)

Matthias: not wanted to re-invent the allele

<Joanne> So, i've not made any progress other than the raw conversion of the warfarin data to RDF

<Joanne> (sorry to have this in-line, but.... all i can do at the moment)

Matthias: in year or so we might want to represent more soph than SNPs, like inversions, deletions and the like.

<Joanne> on the other hand, i have started to read the papers that Elgar got from Michel on the prediction methods for warfarin dosing

Scope of ontology, are we extending to personalized medicine?

Matthias: Might want to re-name ont? There has been some discussion exchange on this.
... major trouble is to avoid modelling too much into basic biology.
... SeqOnt, etc, contain lots and lots of individual things, but lose focus.
... should keep focus on developing treatments. Perspective would help us.
... Protege rich client seems to work pretty well.

+q

scribe: Iker might provide us with individual accounts

Matthias: Agenda: conversion

Joanne: Tim Lebo has been travelling
... looking at papers on predicitve modeling for warfarin dosing

Matthias: Agenda: snpPedia is a gold mine of data, on Semantic Media Wiki
... snpPedia available as RDF; will try to crawl and host
... with some clean-up this might be a very useful dataset
... they are curating mostly by hand

<matthias_samwald1> LDspider

Matthias: next steps, to look into linked data crawlers like LDspider
... Agenda, conversion of other data; should think about other kinds of data that we have not yet discussed
... pretty sure that dbSNP has been RDF-ized
... Michel has already RDF-ized pharmGKB; relation to which ontologies?
... will probably discuss some more. Genotator dataset of gene-disease relationships
... project at Harvard

<matthias_samwald1> http://genotator.hms.harvard.edu/geno/

Matthias: snpPedia, etc should be our primary targets at the moment.
... some version of OMIM is in LOD
... not re-formulate to our ont, just link

Matthas: Sparql endpoint development
... Iker started to create a KB, but not on the call anymore;

Joanne: RPI has enormous expertise in this area; can get help here

Matthias: Want to aggregate all of KB data at a central sparql?

Joanne: But we can federate, need not be all at one place.

Matthias: Yes and no. Tech for federation not yet perfect.

Joanne: Can issue sparql against different sources

Matthias: For demo purposes, works better to have all in one KB
... can use DERI and Michel endpoints
... Sustainability is an issue

Joanne: RPI can sustain across three years for sure

Matthias: Discuss off-line

Fred: Thought about a clinical use case in the Google doc

<Adrien> I don't see the doc on google doc

Fred: deals w interaction between warf, aspirin and plavix, w stent
... use case creates a real-life dilemma for physician
... to prevent cardiac events vs bleeding. An area that is problematic now.

Adrien: Can be useful. We have to be able to use this; if we can succeed here.

Matthias: Realistic and tricky enough to warrant computerized clinical decision support
... complex, should ideally take care of interaction between medications

Fred: If limiting to warfarin only, SNPs are already catalogued
... situation is not complex w warf alone
... interaction with other factors make it more subtle

Matthias: Two levels of soph: 1) system shows a single alert w SNP, plus another alert for drug interactions
... different alerts are separated. But combined risk is not estimated.
... more sophisticated 2) combined risk.
... less soph would leave reasoning to the doctor.
... should we unify alerts and report to doctor?

Fred: Drug interactions already have DBs; situation where physician gets additional info
... but would not want to make a hard-and-fast <determination>
... but make physician aware of genetic risk would be useful

Matthias: Might have to make trade-off. Drug interaction more important than SNPs?
... even if not perfect, sometimes easier is more useful

<Joanne> I need to go. regrets! and thanks!

<Adrien> Bye Joanne

<Joanne> thanks everyone

Matthias: KB should have drug interactions in, as well a pgx
... agenda item to look at drug interactions

<matthias_samwald1> RRAagent, please draft minutes

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: BobP Bob_Powers Fred Joanne Joanne_Luciano Matthas Matthias P22 P50 aaaa aabb adrien ericP https iker matthias_samwald matthias_samwald1
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: 09 Jun 2011
Guessing minutes URL: http://www.w3.org/2011/06/09-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