HCLSIG BioRDF Subgroup/Meetings/2009-04-20 Conference Call

From W3C Wiki
Jump to: navigation, search

Conference Details

  • Date of Call: Monday April 20, 2009
  • Time of Call: 11:00 am Eastern Time
  • 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: ~1 hour
  • Frequency: bi-weekly
  • Convener: Kei Cheung
  • Scribe: Matthias Samwald

Attendees

  • Scott Marshall, Kei Cheung, Jun Zhao, Eric Prud'hommeaux, Matthias Samwald, Rob Frost

Agenda

  • Introduction (Kei)
  • status of the bmc bioinformatics paper (All)
  • HCLS KB update (Matthias, Adrian)
  • Face-to-face meeting: BioRDF breakouts (All)

Minutes

<matthias_samwald> TOPIC --- BMC Bioinformatics Paper

<matthias_samwald> Kei: I have not heard any feedback yet, we are still in the review process

<matthias_samwald> Scott: They extended the review deadline to the 22nd

<matthias_samwald> ...: but there will be late reviews and decision making, so next week will be the earliest for feedback

<matthias_samwald> Kei: I want to thank everyone who contributed to the paper.

<matthias_samwald> Scott: I am enthusiastic about the content in the paper, it is a good look into query federation

<kei> matthias: added new datasets (LODD) to HCLS KB (DERI)

<matthias_samwald> kei: we will also explore linking the Traditional Chinese Medicine (TCM) dataset to the LODD datasets.

<mscottm> http://aida.science.uva.nl:9999/search/

<matthias_samwald> Scott: accessing the HCLS KB at DERI from the AIDA web interface works (via Sesame)

<matthias_samwald> ... our focal point right now is looking at the Gene Ontology graph

<mscottm> http://hcls.deri.org:8080/sparql

<matthias_samwald> ... enter 'http://hcls.deri.org:8080/sparql' into the AIDA thesaurus info field

  • jun (chatzilla@62.82.106.5) has joined #hcls

<matthias_samwald> ... one of the problems we faced was having a large number of named graphs and not knowing what is in them...

<matthias_samwald> ... in addition to the three upper classes of the GO ontologies there are some other things you can ignore

<matthias_samwald> ... when you type in "GABA" you can get autocompletion

<matthias_samwald> ... you can take one of the concepts and drag it to the query to add it, without worrying about namespaces etc.

<matthias_samwald> kei: if i drag two terms into the query builder, is it an "AND" query?

<matthias_samwald> scott: it would default to OR but can be changed to AND

<matthias_samwald> scott: future will bring possibility to name and save queries, and to add more semantics (without SPARQL)

<matthias_samwald> kei: we have developed an application called "Entrez Neuron", a very specific interface for querying neuron-related information. It would be interesting how neuroscientists would use AIDA, a generic browser.

<matthias_samwald> kei: our experience is that users are most interested in textual/tabular forms of information presentation

<matthias_samwald> ... we have to give more thought on how to constraint graph-based transversal in order to make it useful for neuroscientists/biologists.

<matthias_samwald> scott: i agree.

<matthias_samwald> ... but we have to differ between graphs for displaying data and graphs for displaying/editing queries.

<matthias_samwald> eric: about querying... people usually don't use graphs to pose queries when they have become proficient. but there are special cases.

<kei> matthias: graph-based tools for query construction and display but not very useful

<matthias_samwald> kei: depends on the users, bioinformaticians might be motivated, but end-users are not. they prefer a more NLP-based output format.

<matthias_samwald> ... it would be good to have AIDA as one of the project presentations during the BioRDF breakout at the F2F

<matthias_samwald> scott: yes

<matthias_samwald> scott: ... at the moment we cannot access the Allegrograph KB at FU Berlin at the moment (reason: problem with HTTP headers)

<matthias_samwald> TOPIC --- F2F meeting next week

<matthias_samwald> kei, scott, rob and eric will be there, matthias will not be there.

<matthias_samwald> kei: i have been thinking about future directions

<matthias_samwald> ... we have some preliminary results (aTags, query federation)

<matthias_samwald> ... i would like to have some sort of "blueprint" for the future of BioRDF

<matthias_samwald> ... two main goals:

<matthias_samwald> ... e-Science infrastructure (or "cyberinfrastructure")

<matthias_samwald> ... query federation could fit in nicely

<matthias_samwald> ... there is also a social aspect to it

<matthias_samwald> ... we show how to integrate RDF/OWL with social tagging

<matthias_samwald> eric: integrating social and semantic data -- what use cases are there?

<matthias_samwald> kei: e.g. an organ like the brain is researched on different levels, by different institutions. how do we bring them together? this could be an interesting, broad use-case. we could also be more specific (receptors)

<matthias_samwald> eric: relation to Alzforum?

<matthias_samwald> kei: Alzforum focused on discourse, statement level

<matthias_samwald> matthias: novel: decentralized infrastructure

<matthias_samwald> scott: you want to have information dynamically generated

<matthias_samwald> kei: in alzforum, hypotheses can be complementary, related to each other etc.

<matthias_samwald> ... should be discussed at the F2F and even before

<matthias_samwald> scott: there is another use-case for social networks: expert finding.

<matthias_samwald> ... there is a dedicated conference for this topic

<mscottm> Subject: Identifying Researchers on the Biomedical Web (IRBW2009),

<mscottm> Workshop May 13-14

<matthias_samwald> rob: not only finding persons, but also companies (e.g. for collaboration)

<matthias_samwald> eric: there are already some big companies doing this

<matthias_samwald> ... we could potentially look at re-using code from Alzforum

<matthias_samwald> ... the data is curated, so cost is fairly high.

<matthias_samwald> ... statements are not fully formalized in RDF, to my knowledge

<kei> rob: federation over separate repositories (taxonomies) find/connect researchers

<kei> Kei: integration of rdf triples and tags to get people connection?

<kei> scott: atags can be used to associate people with statements

<kei> scott: text mining to do expert finding

<kei> kei: invite bio2rdf (e.g., Michel) to give a talk during the biordf breakout

<kei> scott: ontotext (developer of owlim) joined HCLS IG

<kei> next biordf call will be on May 11.