HCLSIG BioRDF Subgroup/Meetings/2009-03-30 Conference Call

From W3C Wiki
Jump to: navigation, search

Conference Details

  • Date of Call: Monday March 30, 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: Adrian Paschke

Attendees

  • Scott Marshall, Kei Cheung, Jun Zhao, Adrian Paschke, Eric Prud'hommeaux, Matthias Samwald, Michel Dumontier, TN Bhat

Regrets

  • Rob Frost

Agenda

  • HCLS KB update (Matthias, Adrian)
  • Face-to-face meeting (All)
  • BMC Bioinformatics paper submission (All)

Minutes

<apaschke> Topic: HCLS KB update (Matthias and Adrian)

<apaschke> Matthias: worked together with Scott to provide access to AIDA

<mscottm> http://adaptivedisclosure.org/AIDAToolkit/

<mscottm> http://ws.adaptivedisclosure.org/search/

<apaschke> Scott: access W3C HCLS KB (Virtuoso) via AIDA

<apaschke> Scott: web interface - see all the repositories which are available - browse all the repositories and ontologies (hierarchical view of an ontology such as GO)

<apaschke> Scott: visualizes named graphs as possible "repositories"

<apaschke> Scott: search for keywords and add them to the query

<apaschke> Scott: introduces a hierarchy in an ontology - the top level constructs and allows browsing through the hierarchie

<apaschke> Scott: Matthias, do you have plans to fix the second query on the demo page

<apaschke> Matthias: answering the query without named graphs is quite demanding

<apaschke> Matthias: needs optimization

<apaschke> Matthias: needs upgrade to Virtuoso version 6

<apaschke> Matthias: optimization of the memory consumption, etc.

<apaschke> Scott: there seems to be some kind of time limit / time out

<apaschke> Matthias: yes, we introduced it due to too many requests

<apaschke> Eric: shall we wait till the next release of Virtuoso version 6?

<apaschke> Eric: Scott, can you work with the query without named graphs

<apaschke> Matthias: Data is in named graphs - but not used in the current version

<apaschke> Scott: tried to load UMLS into Sesame - required a 64bit machine

<apaschke> Scott: will try with a Amazon cloud

<apaschke> Scott: have to figure out how to split costs, e.g. by using a EC2 could

<apaschke> Yun: we currently run a 64 bit server in our group for FlyBase

<apaschke> Jun: currently funded by our group

<apaschke> Jun: costs proportional to the size of the triple store and the number accesses

<apaschke> AllegroGraph version 3.2 release, query optimizations & more

<apaschke> average, an order of magnitude faster

<apaschke> performance improvements include RDFS++ Reasoning, get-triples, client/server bulk loading, SPARQL, Jena and Sesame

<apaschke> Gruff graph-based browser for AllegroGraph

<kei> scott: contacted a staff member about issusing sparql through endpoint, but there is a security concern ...

<apaschke> Scott: might try to add a new layer of interfaces for AIDA to access the AllegroGraph store

<apaschke> Matthias: the DERI Virtuso store is also accessed via a http SPARQL interface by AIDA

<apaschke> Matthias: should also work for the instance of AllegroGraph

<apaschke> Scott: Will look into that again

<apaschke> Gruff graph-based browser for AllegroGraph - a grapher that displays information on data stored in AllegroGraph in a variety of methods

<apaschke> AGWebView - A Web Browser Server for AllegroGraph - AGWebView gives you access to your triple store data via from an ordinary Web browser. Available for Windows, Linux, and Mac OSX.

<kei> when these are set up, they can be demo'ed to the group.

<apaschke> Topic: * Face-to-face meeting (All)

<apaschke> kei: F2F meeting end of April in Boston

<apaschke> kei: what will the BioRDF group present?

<ericP> -> http://esw.w3.org/topic/HCLSIG/Meetings/2009-04-30_F2F F2F home page

<apaschke> kei: breakout sessions for BioRDF; who will attend

<ericP> ericP

<scott> will attend

<michel> i intend to

<matthias> may be depending on funding situation, but a DERI representative will attend

<apaschke> Kei: present dynamic queries, aTag effort, add new activities, collaborations with other task forces (e.g. LODD, Jun's herbal data set ),

<apaschke> ericP: there might be a possibility to dial-in

<ericP> i stuck an anchor on http://esw.w3.org/topic/HCLSIG/Meetings/2009-04-30_F2F#BioRDF so we can nav there directoy

<apaschke> Kei: ... collaboration with Bio2RDF

<apaschke> Eric: would like to encourage people to add ideas

<apaschke> Scott: federated queries using more bio-informatics data in the demo

<apaschke> Topic: BMC Bionformatics paper submission

<apaschke> Kei: about one week left

<apaschke> Kei: deadline, April 6th

<apaschke> Kei: main part is federation

<apaschke> EricP: built a mapping file and a small database to query it

<apaschke> EricP: little bug in the encoding site, currently crashes parser

<apaschke> EricP: spread sheet or large databases ?

<apaschke> EricP: make the table structure flat or nested / linked?

<apaschke> Kei: biomedical domain data often disseminated in spreadsheet format -> i.e. flat

<apaschke> Kei: could be presented as simple RDF graph

<apaschke> Kei: leightweight ontologies built from flat rel. schemas

<apaschke> Kei: could be interesting to show this mapping

<apaschke> EricP: explain mapping file for both structures (flat and deep database schema)

<apaschke> Scott: seems to be useful for the reader

<apaschke> EricP: gene records with different types; point to other tabels -> is a realistic representation

<apaschke> EricP: could show the two mapping approaches in the paper

<apaschke> Matthias; would only show the first mapping and simply state that we could also do the other approach (without explanation)

<apaschke> Kei: could a compromise - show the simple mapping in details and simply discuss the other approach in the discussion section

<apaschke> Scott: add an example in the discussion section

<apaschke> EricP: there wouldn't be much difference between the two, except some rules which can be explained

<apaschke> EricP: shall we discuss on the public HCLS mailing list?

<apaschke> Scott: in general yes. But 500 subscribers might wonder - since it is a discussion about the contents for a paper

<apaschke> EricP: send some general discussion about technical issues (not details of the paper)

<apaschke> Kei: DBPedia, receptor table "B"

<apaschke> Kei: federate triple formatted sources

<apaschke> Matthias: integrate statements which use rdf-a tags in some of the triples

<mscottm> Zakim, this will be hcls

<Zakim> ok, mscottm, I see SW_HCLS(BioRDF)11:00AM already started

<apaschke> EricP: no parser for RDFa currently

<matthias_samwald> http://hcls.deri.org/atag-data/pastebin.html

<apaschke> Matthias: XSLT to translate RDFa via GRRDL to RDF/XML

<matthias_samwald> http://triplr.org/rdf/http://hcls.deri.org/atag-data/pastebin.html

<matthias_samwald> (gives you the RDF/XML version of the aTag RDFa pastebin)

<ericP> Human EG for 2550 9568

<ericP> Rat: 81657 83633

<ericP> Mouse: 54393 242425

<matthias_samwald> http://en.wikipedia.org/wiki/GABAB_receptor mentions "2550"

<ericP> excellent. and the DBpedia query which will retrieve that?

<apaschke> Loom would be certainly interesting in this context too http://www.loomp.org/

<matthias_samwald> yes.

<apaschke> Kei: future directions - open query federation

<apaschke> Kei: exploit meta data to open query federation

<apaschke> Kei: link together data

<apaschke> Kei: URI issues in query federation

<matthias_samwald> wow, loomp looks a lot like aTags :)

<apaschke> Scott: how can we proceed in the next days?

<apaschke> Kei: by email

<apaschke> EricP: by IRC?

<apaschke> Scott: use HCSL IRC? might be taken

<apaschke> EricP: hcls is fine

<apaschke> Kei: let's use this IRC to proceed and use emails

<apaschke> Kei: let's have a working BioRDF call next week just on the paper

<apaschke> Kei: our next regular BioRDF call will be in two weeks (April 13)