<michel> /invite zakim #HCLS
<mscottm2> having trouble dialing in with skype again - looking into 'sip' clients..
-> http://www.w3.org/2006/tools/wiki/Zakim-SIP Zakim-SIP
<michel> michel: nigam @ stanford, working with clinical notes; using NCBO annotator, odds testing between drugs and adverse events
<michel> ericp: can we link that output with semantic web?
<michel> matthias: aTags to markup sentences, adds knowledge
<michel> ericp: blobs of text such as clinical encounters, nlp -> tag clouds; drug production pipeline? safety/efficacy test
michel: visiting professor is
investigating how ontologies can be used to aid
navigation
... i.e. get from A to B in fewer clicks
... we could ask how a tag cloud could help
<michel> ericp: having a tag soup is not all that useful ; but having structure might be
<michel> ... can it be used for CDS ?
<michel> mscottm2: ncbo annotator is an entity recognizer; breast cancer use case where labelling is useful
<Bosse> Predicate mappings in LinkedLifeData, http://linkedlifedata.com/sources
<michel> ericp: imagine we had super structured information about these patients - we came up with question we could ask. imagine we had structured events, but only as tag clouds - would the use cases operate in that space
<michel> mscottm2: bag of text could be used to generate a semantic neighbourhood... could also project against ontologies to see what is being talked about
<michel> bosse: causal relation extraction together with hypotheses to try to explain those questions
<michel> ... the tag cloud could be used to help explain a phenomena
<michel> ericP: something to help keep track of assertions
<michel> bosse: if you had a knowledge base where you could annotate and share, it could be very useful
<michel> ... if we could find a community with an interesting enough problem, then the kb could serve as a real resource
<michel> ericP: recording institutional knowledge has to immediately benefit, or has to have a big stick
michel: some pharmas and other
organizations track contributions towards a patent
... your motivates is more info that can be used for a product
or servce
... is your contribution information that fills a gap and leads
to inference?
... that would be a way to measure contributions
... you could pose a question a priori to enhance this
metric
... could ask SemWeb folks at elsevier (Alan Yagoda, Sigram,
Iker)
... Alan raised the issue of data quality accuracy
... SIDER was based on matching product labels, leading to
inaccuracies
... issue: freshness of data
<michel> ericp: are there examples of pharma, do record extra information to help other groups - is it a cultural norm?
<michel> bosse: we could be helped a lot more in content authoring
<michel> ericp: need grant to develop an open source project
michel: Bosse said we have to
find the community and the problem
... it would be hard to find funding unless it came from a
VC
... for any other funding org, we'd have to present the problem
and show how a SemWeb tool is the solution
mscottm2: LODD Note is waiting
for a couple edits, then moving to HTML (1 week)
... we've identified a couple things which need fixing in the
expression study note
... we need alpha testers to read the doc and adapt tools to
reflect the note
... also need to adapt Jim Mckuster's R script or find
equivalent
mscottm2: we discussed John's
work saturation and stepping down as chair
... decided that we should reach out to IHE (International
Health E) contacts
... IHE is for clinical IT. includes imaging
<mscottm2> Here's a link for IHE (about the upcoming Connectathon in Bern in May): http://ihe-europe.net/connectathon/133/cat-2012
michel: John wanted to move away from terminology to applications
matthias: will have first meeting
on Thursday
... am expanding our current data to meet pharmacogenomics use
cases
<mscottm2> Another relevant IHE link: http://www.dcm4che.org/confluence/display/ee2/IHE+Integration+Statement
matthias: next step is to see how we can use SPARQL, SPIN and OWL to make context-senstitive decision support algorithms for e.g. reminders, warnings, recommendations
<scribe> scribenick: ericP
mscottm2: exploratory discussion
with lots of enthusiastic folks
... would need to consult minutes to give accurate summary
michel: we've been looking at
clinical annotations in pharmgkb
... can be released for research (not publication) with
contract
... added links to DBSNP
... challenge is to extract and get access to portions of
DBSNP
mscottm2: we've discussed I2B2 or
SAGE Bionetworks datasets
... would be good to create an inventory
michel: look for main wiki page link to "potential data sources"
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) Found ScribeNick: ericP Inferring Scribes: ericP WARNING: No "Present: ... " found! Possibly Present: Bosse HCLS MacTed P16 P5 Scott_Marshall aaaa aabb egonw ericp joined matthias matthias_samwald michel mscottm2 scribenick You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 Guessing minutes URL: http://www.w3.org/2012/03/06-HCLS-minutes.html People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]