HCLSIG/LODD/Meetings/2008-10-29 Conference Call

Conference Details

• Date of Call: Wednesday October 29, 2008
• Time of Call: 11:00am Eastern Daylight Time (EDT), 15:00 Greenwich Mean Time (GMT), 16:00 Central European Time (CET)
• Dial-In #: +1.617.761.6200 (Cambridge, MA)
• Dial-In #: +33.4.89.06.34.99 (Nice, France)
• Dial-In #: +44.117.370.6152 (Bristol, UK)
• Participant Access Code: 4257 ("HCLS").
• IRC Channel: irc.w3.org port 6665 channel #HCLS (see W3C IRC page for details, or see Web IRC)
• Duration: ~1h
• Convener: Susie

Agenda

• Overview of LarKC - Bosse
• Further development of the use case - Bosse, Susie
• Further exploration of the data sets - Anja
• AOB

Minutes

Attendees: Anja, Kei, Jun, EricN, Susie. Bosse, Vassil, EricP

Scribe: Susie

<sstephen> Bosse presents on LarKC

<sstephen> http://esw.w3.org/topic/HCLSIG/LODD/Meetings/2008-10-29_Conference_Call -> LarKC slides

<sstephen> Kei: Will you include environmental data?

<sstephen> We haven't incorporated environmental data, but it would be interesting to do so

<sstephen> Incorporate full data sets

<sstephen> Can test performance yourself at http://linkedlifedata.com

<sstephen> It's an integrated store - which is the most scalable approach

<vassil> The integrated approach supports efficient reasoning, but limited scalability; the current scalability limit is several billions of statements

<sstephen> All SPARQL queries can be answered in 10 seconds

<vassil> A list of predefined complex demonstration queries could be answered in 10 seconds. (e.g., a query to get all the data from the repository will take much more time only to transfer the amount of data)

<sstephen> Will moved to a distributed architecture with time

<vassil> As next step is to move from OWLIM to LarKC platform which will use distributed architecture

<vassil> Currently 403,361,589 explicitly asserted statements to generate 1,159,857,602 statements including inferred by the reasoner

<sstephen> Next step is to focus on tractable reasoning, incorporate additional information, make sure query performance is good

<vassil> - Next step is to extend the knowledge base with additional information (we currently use tractable reasoning - OWL-Horst)

<sstephen> Move onto information extraction with time

<sstephen> Project incorporates engineers who are looking to build a scalable reasoning engine

<sstephen> Focus on platform, rather than becoming sole platform for life sciences

<vassil> Focus over engineering process of implementation of scalable reasoning platform to support RDF, rather than development of universal RDF representation of lie science data sources

<sstephen> Would like us to test out the system

<sstephen> Business case for LODD -> http://esw.w3.org/topic/HCLSIG/LODD/Business

<sstephen> Want to keep clinicaltrial.gov at the center, which does have an XMl representation

<sstephen> Want to keep focus on patients, physicians, and pharma

<sstephen> Need to identify starting point for the demo

<sstephen> Anja gives updates on data sets

<sstephen> Oktie has RDF version of clinicaltrials.gov

<sstephen> Has initial version of the LODD cloud

<AnjaJentzsch> linkedct.org

<sstephen> Anja: linkedct.org is a great starting point for the demo

<sstephen> Anja: hope to have more by next call

<sstephen> Susie proposes focus on AD, although demo would incorporate all data so examples could cover other TA

<sstephen> Kei: wants to incorporate the LODD work into queries that the BioRDF group performs

<sstephen> Progress on action items

<sstephen> Susie & Bosse have done further work to define the use case

<sstephen> Anja has assessed the data sets in the Wacholder paper and will add information to the wiki

<sstephen> Anja has created a LODD cloud and will post onto the wiki

<sstephen> Oktie has confirmed that we can use IBM's RDF version of clinicaltrials.gov

<sstephen> Chris has invited Bio2RDF to present during an upcoming call