<michel> http://informatics.mayo.edu/LCD/images/c/c6/LCDatHCLS.ppt

<sivaram> Hi This is Sivaram Arabandi and Alan Yagoda from Elsevier

http://informatics.mayo.edu/LCD/index.php/Main_Page

Jyotishman Pathak: The Linked Clincal Data Project, from Mayo Clinic

scribenick bobP

Jyotishman: Have been involved in HCLS, Clinical interoperability
... started last fall, up and running since early this year
... Mayo has patient data since 1907
... focussing on chronic diseases, but also risk for carviovascular diseases as well
... LCD architecture using Virtuoso
... MCLSS endpoint, exposed and made queryable
... private sparql endpoint will stay within Mayo firewall
... Virtuoso commercial edition, discussion
... have used the TMO! plus added a few things, from a cancer ontology
... drugs, diagnosis info, lab measurements, demo info
... queried sider at Stonybrook, created federated query
... an early proof-of-concept w real clinical data, learning experience
... starting to look at linked-data api
... IT likes idea of leveraging linked data, but not going to write sparql
... need more user-friendly graph UI

<Lena> sounds interesting, looking forward to hearing the rest :)

Matthias: Use of federated query, conformant?

Jyotishman: Also looked at topbraid and twinkle(?)
... audience mostly clinical investigators; Don't quite get it yet
... needed to convince people

(?Elsevier): Are endpoints stable?

Jyotishman: Have played around w sider and Stonybrook.
... mixed results. Endpoints are up and down, so considered put everything local

<matthias_samwald> ?Elsevier = Alan Yagoda

Jyotishman: but still want to make public, w institutional challenges
... also struggle to find endpoints; 3 endpoints for RxNorm
... 3rd issue, data may be old in public endpoints
... have not figured it all out yet, re endpoints.

<Lena> rxnorm has a Java API: we should instead of converting to RDF, place a SPARQL endpoint on top of that

<Lena> how doable would that be? (similar discussion for other SPARQL endpoints)

Jyotishman: local solution, personal OP does not really leverage open linked data

Alan Yagoda: Lots of endpoints are stale, not reliable. Have started to pull some data in

<iker> +1 Alan opinion, SPARQL endpoint are not reliable

Jyotishman: Also need to deal w issues of latest update of RxNorm

<mscottm2> Relevant links: http://labs.mondeca.com/sparqlEndpointsStatus/index.html

<mscottm2> http://thedatahub.org/

<iker> Please see the speaker queue

Matthias: Important also is to have ways to convert legacy data; Alan +1

Iker: Q

<matthias_samwald> (sorry for ignoring the queue!)

<mscottm2> no prob

Jyotishman: Project in google code, so API is going public, do not want to rebuild what others have done

<xorcon> http://code.google.com/p/linked-data-api/

<Lena> (don't agree with RDF-izing... we should be talking about drinking from the source!)

Iker: (Should create your own custom scripts)
... commercial environment, you are responsible for quality of data!

Jyotishman: RxNorm weekly update, we pull a virtual RDF graph

Scott: Q Federating endpoints. Virtuoso, need to use select-as structure

<iker> I must attend another meeting and will leave now, Jyotishman really great project!!! congratulations!!!

Jyotishman: Using service keywords, problems.

<mscottm2> Sorry Bob - that was hard to scribe - it was a comment about Virtuoso not supporting 'BIND' from SPARQL 1.1 - probably doesn't apply to the problem they were having

<trish_whetzel> is there a use case for translating and adding the weekly rxnorm updates as RDF to a sparql endpoint?

(?): (details re Virtuoso tech here)

Scott: Fed queries and reliability: have been hearing this for quite a long time
... pharma has found sparql going down all the time
... LinkedCT improved, updated nightly
... NLP -> linked data not quite good, going out of semweb
... make a local copy (in real-time) when the public sources goes down
... notion that there is one unique copy is going to change
... different copies have different styles of RDF
... culture *will* change, publish your own copy if you like it

Jyotishman: Yes, this may be the approach we should take
... BobF here are Mayo, have been looking at SNPedia

<mscottm2> q

Jyotishman: Goals: navigate EMR w OMIM and dbSNP, look at co-morbidities
... did not find a good dbSNP endpoint
... did dump of dbSNP; BobF wrote script to filter out;
... took brute force approach to download, etc

<Lena> we can host dbSNP here at DERI

<michel> Bio2RDF can host as well

<xorcon> paper on PhewAS: http://www.ncbi.nlm.nih.gov/pubmed/20335276

Sivaram from Elsevier: Q How are you actually linked the terms?

<mscottm2> Your voice fell out for me?

<matthias_samwald> it works for me

Jyotishman: Queries w sider and RxNorm, have been using Rx terms

<michel> works for me

<michel> check identifiers.org

Jyotishman: point is good re vocabulary

<trish_whetzel> are the needed mappings stored in bioportal?

Sivaram: Connecting genomic data to EMR, quite a wide domain

Jyotishman: Some data already have the snomed, but not exosed yet by endpoints
... starting to run queries in earnest just now

Matthias: Q snpedia, using rdf version?

Jyotishman: They will not give us access. Not an endpoint yet for snpedia
... Matthias: SNPedia producing only part of info that you will need

Matthias: There are other datasets, I can point you
... Q expanding TMO; collaborate on TMO development?

Jyotishman: Yes. Extended only of necessity. Definitely interested in pursuing collaboration

<trish_whetzel> is there a use case for translating and adding the weekly rxnorm updates as RDF to a sparql endpoint?

<trish_whetzel> are the needed mappings stored in bioportal?

Lena: Should focus on pushing providers to give either sparql or sql endpoints, at the source

<mscottm2> See also NCBO's upcoming SPARQL endpoint: http://alphasparql.bioontology.org/

Lena: danger is lack of funding; data may not be in synch
... ask sources to build the endpoints

Jyotishman: Example Sider, OMIM endpoints would be fantastic
... +1

Scott: Republish w metadata *about* your local copy

Lena: Semweb 10 yrs. Most providers may actually be willing to do so

Jyotishman: A about BioPortal/RxNorm. BioPortal does not map Sider/RxNorm

<matthias_samwald> RxNorm is in BioPortal: http://bioportal.bioontology.org/ontologies/1423

<trish_whetzel> thanks - was referring to term mappings.

<matthias_samwald> (via UMLS)

<trish_whetzel> bioportal does include rxnorm via the mayo lex backend architecture

Jyotishman: Thank you on behalf of Rick and BobF here at Mayo.

<Lena> RXnorm in bioportal does not seem to have instance data, is that right?

Michel: Interested in dbSNP here, maybe discuss about putting into public domain

<sivaram> Thanks Jyoti - on behalf of Sivaram and Alan

<mscottm2> Thanks Jyoti!

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 Jyotishman JyotishmanPathak Lena Matthias Michel P1 P13 P14 Scott Scott_Bauer aa aaaa aabb aacc bobP epichler ericP iker matthias_samwald mscottm mscottm2 rkiefer simona sivaram trish_whetzel xorcon
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

Got date from IRC log name: 27 Oct 2011
Guessing minutes URL: http://www.w3.org/2011/10/27-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