HCLSIG BioRDF Subgroup/Meetings/2010/10-25 Conference Call
Conference Details
- Date of Call: Monday, October 25, 2010
- Time of Call: 11:00 am Eastern Time (5 pm CET)
- Dial-In #: +1.617.761.6200 (Cambridge, MA)
- Dial-In #: +33.4.26.46.79.03 (Nice, France)
- Dial-In #: +44.203.318.0479 (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), Quick Start: Use http://www.mibbit.com/chat/?server=irc.w3.org:6665&channel=%23hcls for IRC access.
- Duration: ~1 hour
- Convener: M. Scott Marshall
- Scribe: Lena Deus
Participants
Scott Marshall, Kei Cheung, Lena Deus, Rob Frost, Michael Miller, Eric Prud'hommeaux, Matthias Samwald,
Agenda
- Introduction
- Next phase of BioRDF
Minutes
<Lena> mscottm: fuuture bioRDF focus is on use cases in the neurosciences domains, federation and provenance <Lena> mscottm: by working with people in the community (EBI, etc), we can setup standard services for microarray data <Lena> Linked open data mailing list is discussing best practices, about, among other things, provenance metadata <Lena> role of bioRDF in LOD is to provide federation use cases <Lena> (ex: what are the requirements for provenance?) * michael has joined #hcls <Lena> pharmGKB lexicon in RDF dataset with NCBO SPARQL endpoint <Zakim> + +1.206.605.aabb <Lena> can we integrate that data with what we already have in bioRDF? <Lena> cooperate with consumers of EBI microarray data <Lena> see if there is something that EBI has that we can learn from <Lena> can we increase federability of our knowledge resources (i.e. HCLS Kb)? <Lena> if a URI is used to refer to a repository, things can be said about the graph itself <Lena> reflective statements - added to the graph; if there is provenance metadata, it should be using the graph URI <Lena> the idea is to use graph URI as they are used by EBI <Lena> things like rdfs:label can be used to refer to the graph URI <Lena> void can be used to talk about each of the grpahs in the Kb <Lena> such as the ones we have used in the example queries <Lena> kei: how to use rdfs:label to refer, for example, to a pathway graph? <Lena> mscottm: when NCBO provides a lengthy, complete, URI; to display it in the interface, the label can be used <Lena> that way, the labels can be queried <Lena> string matching through labels as part of the search <Lena> what if that was a type instead? <Lena> ericP: that way, hierarquies can be added <ericP> { <...g218> a hcls:PathwayData . hcls:PathwayData rdfs:label "Pathway Data" . } <...g218> { ... pathway stuff ... } <Lena> mscottm: the idea is to enable developers to stay in SPARQL to build the datasets <Lena> mscottm: in the case of disease ontology, by creating this sets of provenance, we can 1) query the sets directly (use the information in the provenance to build the query itself) <Lena> the hardest work is deciding on the predicates to use <rfrost> Got to drop off early - talk to everyone next call <Lena> create a description using widely used rdf terms to create a vocabulariy <Zakim> -rfrost <Lena> ericP: discoverability and composability <Lena> mscottm: a lot in common with SADI <Lena> mscottm: SADI - given an rdf graph describing the input and some rdf neighbourhood <Lena> ericP: SADI unit of charatecterizarion of services is al at the predicate level <Lena> ericP: this makes SADI look much more like voiD <Lena> ericP: at NCI, some work in this space also <Lena> ericP: integrating web services using SPARQL (code wrote a while ago) <Lena> ericP: may be able to make a better use of stuff at NCBO <Lena> (correction - no NCBI, but NCBO above) <Lena> ericP: encourage NCBO to do SPARQL such that it is more like a service than a function <Lena> ericP: plan is to integrate with stuff at NCBO, TMO and bioRDF (next 3 months) <Lena> ericP: in a function, inputs can be left blank whereas in service they cannot <Lena> ericP: execution path affects the outcome <Lena> mscottm: microarray / NCBO link <Lena> kei: NCBO implemented SPARQL endpoint; <Lena> ericP: service constrains in SPARQL do not allow to specify default named graph <Lena> ericP: we needed to add a default graph parameter <Lena> ericP: are there going to be a lot of endpoints where we need to specify the named graphs? <Lena> kei: e.g. in database systems querying, we can specify the db name as part of the query <Lena> kei: that is often quite useful; along the same lines, that parallel in SPARQL would be useful <Lena> ericP: the parallel is informative, does not tell us if that funcionality if fully required <Lena> ericP: a SQL command can do that in a database; the required databases would have to be loaded first <kei> lena: use the same disease id to link diseome dataset with other other sets <kei> sorry, I have to go to another meeting now... <Zakim> -Kei_Cheung * kei has quit IRC (Quit: kei) <mscottm> I have to go to.. <mscottm> I have to go too.. <Zakim> -matthias_samwald <Zakim> -mscottm * jodi has joined #HCLS * jodi has quit IRC (Quit: jodi) * michael has left #hcls * jodi has joined #HCLS * jodi has left #HCLS <ericP> SPDL * matthias_samwald has quit IRC (Quit: Leaving.) <ericP> http://www.w3.org/2007/Talks/0511-egp-spdl/#%281%29 * jodi_ has joined #HCLS <Zakim> - +1.832.386.aaaa <Lena> my connection fell.. :( <Lena> but I can't call back <Lena> that's ok, I have to go in a few minutes * ericP sorry Lena! <Zakim> -EricP <Zakim> - +1.206.605.aabb <Zakim> SW_HCLS(BioRDF)11:00AM has ended <Zakim> Attendees were mscottm, Kei_Cheung, EricP, +1.832.386.aaaa, rfrost, matthias_samwald, +1.206.605.aabb * ericP also sorry that i didn't notice before