See also: IRC log
<matthias_samwald> thanks
<scribe> scribe: ericP
agenda linke?
link?
<Jun> http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Meetings/2009-01-12
<kei> http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup
kei: updated BioRDF homepage this
weekend
... wrote introductory paragraph
... please make suggestions on wording
... listed objectives:
... .. enhance HCLS KB; new datasets to help KB span HC and
LS
... .. identify scientic use cases
... .. maintainability of data sets
... .. disseminate the HCLS KB (and SemWeb) gospel
Jun: will we include the pointers to the new KBs query interfaces?
kei: good idea
Jun: relevent literature, e.g. special issues on life science data integration
kei: could go under meetings. if the list gets to long, could move to a separate page
<scribe> ACTION: kei to start a list of relevent literature on the BioRDF homepage [recorded in http://www.w3.org/2009/01/12-hcls-minutes.html#action01]
matthias_samwald: [update on HCLS
KB at DERI]
... SPARQL endpoint up and running
... data consists only of neurocommons
... not finalized a DRUPAL interface
... looking at performance integrations, e.g. reconfiguring
index to improve performance of queries which don't include
named graphs
kei: do you plan to incorporate additional sources?
matthias_samwald: within BioSioc,
yes. but need discussion within the BioRDF group to see what we
want in there
... suggested we take a closer look at what is coming out of
Virtuoso
... neurocommons KB will be offered in the Amazon cloud
computing infrastructure
... can be used as a comodity by us
<AdrianP> http://www.corporate-semantic-web.de/hcls.html
AdrianP: we have an AllegroGraph
triple store at Free Univ Berlin
... have loaded the Senselab data
kei: senselab ontology available in the neurcommons KB? so it's redundant?
AdrianP: yes
kei: we're not trying to compare
the performance of these two triple stores
... but is good to make folks aware of the features available
in the two
matthias_samwald: [update on SWAN SIOC]
<AdrianP> composition of multiple triple-stores into a single virtual store
RobFrost: triple store comparison
helps you learn about and select proprietary features
... but certainly having non-overlapping content is the driving
force
... [DBPedia and neuro-receptor use case]
kei: need to look at what
component needs to be added to one of these interfaces
... can we identify a couple queries that are useful for
comparison purposes, while we are thinking about the dataset to
include
RobFrost: tying in another, non-HCLS database, including less-structured wiki content seems intriguing
<AdrianP> AllegroGraph supports RDFS++ Reasoning
<AdrianP> and integrates with RacerPro for DL reasoning
AdrianP: AllegroGraph has query support for RDFS++
kei: can is support this construction at arbitrary levels?
mscottm: two ways to reach the
same result:
... .. e.g. go off-line with Pellet, run the transitive
closures, and re-insert into the graph
... .. query-time reasoning is of course more convenient
<AdrianP> yes, we need to try it out
mscottm: there are some stores
that are supposed to reason over a 200M triples
... many tasks don't scale to that level
<matthias_samwald> yes it has
<AdrianP> yes
<matthias_samwald> (meaning: virtuoso has features for subclass/suproperty reasoning)
<Jun> http://docs.openlinksw.com/virtuoso/rdfsparqlrule.html
s/\?//
<RobFrost> I can test out the virtuoso subsumption support on the yago receptor tree in dbpedia
matthias_samwald: [update on
BioSioc]
... email list and a rudimentary wiki page
<RobFrost> http://dbpedia.org/class/yago/Receptor105608868
matthias_samwald: will give a
short presentation next week
... idea is to take an approach simpler than the HCLS KB
... take only the important statements
... plan to optimize for querying
... and develop UIs
... expect to use RDFa to embed SIOC statements
... want to explore the A tag for representing [structured]
biological statements in [unstructured] pages
... thesis: can be useful given intuitive interfaces
<matthias_samwald> "aTag", a shorthand for "associative tag".
<kei> http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/QueryFederation
<matthias_samwald> I need to leave now, bye!
kei: we need to demo queries
across heterogeneous DB
... first is the neuroscience data integration use case
... in the last charter, we showed how SemWeb could be used for
integrating life science data
... want to re-use that data
... have drilled down a little bit in thinking about linking
neurons cells and genes
... need to consider receptors
... can combine receptor trees from multiple sources
... once folks find a receptor of interest, users can drill
down
http://www.w3.org/2004/10/04-pharmaFederate/
kei: it seems we have diverse
approaches to federation
... how can we coordinate these?
mscottm: one prob is "what part
of the federation system has what's interesting to me?"
... the common naming system makes this a bit easier
... i.e. i can register myself as supplying information about
X
<RobFrost> http://dbpedia.org/page/5-HT3_receptor
<AdrianP> bye
This is scribe.perl Revision: 1.133 of Date: 2008/01/18 18:48:51 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) FAILED: s/\?// Found Scribe: ericP Inferring ScribeNick: ericP Default Present: RobFrost, Kei_Cheung, EricP, Jun, mscottm, matthias_samwald, AdrianP Present: RobFrost Kei_Cheung EricP Jun mscottm matthias_samwald AdrianP Got date from IRC log name: 12 Jan 2009 Guessing minutes URL: http://www.w3.org/2009/01/12-hcls-minutes.html People with action items: kei[End of scribe.perl diagnostic output]