HCLS Telecon

 

 

<matthias_samwald> hi, i am only listening (sitting in a public place)

i have pretty bad static as well

introductions

KevinDoyle: i'm from Teranode
... we have a product that uses Oracle's triple store
... i serve as an interface between customer (so i know their applications) and developers

flyweb [Jun]

<Jun> flyweb

<mscottm> http://users.ox.ac.uk/~zool0770/presentations/2008_11-13_HCLS_Call.pdf

<Jun> http://openflydata.org/flyui/build/apps/imagemashup2/

<matthias_samwald> i see both

mscottm: you have a lookup for overloaded gene names. is that local?
... so you have all synonyms available?

Jun: yes, quite authoritative within the Dres community

slide: [SPARQL queries]

jun: sparql tools with 8M [triples? records?] were too slow on case-insensitive searches so we down-cased everything

slide: [The RDF data sources]

Jun: we used d2rq
... we were conservative so we could impose minimal semantic interpretations

<Jun> chris mugall

<Jun> mungall

Jun: Chris Mungall published one with much more semantic interpreation

slide: [SPARQLite Protocol]

Jun: streaming was very important to our app
... otherwise we ran into memory limitations
... we want to eliminate killer queries (including CONSTRUCT)

slide: [Lessons]

Jun: we find the unified RDF approach easier, but saw no overwhelming reason we couldn't use conventional tooling
... but RDF proved easier when we had to connect to another data source (an RDB)

<Jun> unbounded

OPTIONAL { ... ?x } FILTER (!BOUND(?x))

ericP: was streaming sparql getting around a protocol limitation, or an implementation limitation

jun: implementation

ericP: what were your killer queries?

jun: will send

mscottm: what's the cost and challenge to cloud computing for this?
... you said text matching was expensive
... e.g. identifier completion?

<matthias_samwald> Amazon EC2 pricing: http://aws.amazon.com/ec2/#pricing

jun: we don't want to do too many LIKE queries -- leave that to user
... text completions is lower priority than adding data sources to flyweb

kei: is the use case focused on developmental biology, and will you expand the biological domain?

jun: yes, and we are looking for opportunities to do so

kei: are you considering other gene expression data, e.g. microarray?

jun: yes, the @@X and @@Y databases have microarray data

<Jun> flyatlas

jun: we'll have some microarray data
... for instance, we'll integrate flyatlas into the next release of flyweb

<mscottm> http://www.w3.org/2001/sw/hcls/

f2f [mscottm]

mscottm: f2f was at the W3C TPAC in mandelieu FR. allowed us to geek with other folks

<matthias_samwald> someone needs to mute his/her phone.

mscottm: break-out session focused on integrating Bio2RDF with the LODD project

-> http://esw.w3.org/topic/HCLSIG/F2F/2008-10_F2F HCLS IG face2face agenda

Summary of ISWC [ericP]

[scribe was speaking]

Outreach Update [SusieS]

Susie: SemTech conf is accepting abstracts for spring 2009

-> http://www.semantic-conference.com/ SemTech

Susie: if we get enough proposals, we could get our own track
... i'll be a cochair for bio ontologies sig at ISMB next year
... interested in improvement feedback
... working with CDISC and HL7