HCLSIG/Meetings/2009-10-01 Conference Call
- Date of Call: Thursday October 1, 2009
- Time of Call: 11:00am Eastern Daylight Time (EDT), 16:00 British Summer Time (BST), 17:00 Central European Time (CET)
- 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: ~1h
- Convener: Scott
- Scribe: TBD
- News Updates by BioRDF, COI, LODD, Pharma Ontology, Scientific Discourse, Terminology - All
- Next F2F - All
- Collecting use cases for new SPARQL features
- Provenance Incubator Group - Scott
- SKOS testimonial from HCLS
- Semantic Web Applications and Tools for the Life Sciences Workshop, Abstracts due: Oct. 5 - Scott
W3C HCLS IG 01 Oct 2009 Attendees
Anja, Matthias, Scott, Oktie, Egon, EricP, Michel, AllenHobbs, Elgar, Bosse, Julia
- BioRDF update
- Translational Medicine
- scientific discourse
- skos testimonial
- Summary of Action Items
<egonw> there is no way to call in via plain Skype, is there?
<matthias_samwald> ack ??P9
<michel> egon, i am calling by skype...
<michel> to the US number
<egonw> michel: to what account? or skype out?
<egonw> ok, latter
<michel> skype out
<mscottm> Matthias: It is tricky to export all the named graphs
scribenic: ericP BioRDF update
matthias_samwald: i will send mail with an update
mscottm: you said there were problems with named graphs?
matthias_samwald: yes, difficult to dump named graphs ... various work-around, but virtuoso runs out of memory... ... neurocommons import scripts work only with virtuoso 5 ... openlink can't help 'cause we keep running out of memory
mscottm: you now have more RAM? disk?
matthias_samwald: more discs, improves query performance
mscottm: how many triples do we have, and how big can we get? ... e.g. 2B like for LOD cloud
matthias_samwald: 350-400M now ... expect it would be difficult to hit billions with one machine
mscottm: how much ram do you have?
matthias_samwald: 8G, i think
mscottm: neurocommons was around 10x faster than old deri
matthias_samwald: they had a large dedicated server. may have changed to amazon ... am considering a dev KB and a more stable KB ... so you can pick the appropriate endpoint ... if we keep updating the KB, it can confused HCLS users
ericP: how does the LOD cloud hit 2B? can we do that?
matthias_samwald: probably a big machine of cluster, plus expertise of openlink
matthias_samwald: do we need 2B or are we just impressed by their KB? ... working on a loose but ongoing update mechanism ... LOD is just a dump of everything out there
mscottm: there are synonymous identifiers in LOD ... if we see shared-names, or other logical unifying names, we can use them ... that would be a good goal for LODD ... not just large DBs but tightly-coupled DBs LODD
mscottm: congrats on iTriplication challenge!
AnjaJentzsch: tx, and tx to everyone who contributed to iTrip
<mscottm> ..focusing on how to make linked metadata
AnjaJentzsch: Jun Zao , Bosse and I are currently working on how to assert confidence links and links to the tools used to extract the triples ... currently setting up a relation-finder instance ... finds relations between two instances ... e.g. see how aspirin and headaches are generated
mscottm: +1 for the relation finder ... can you use the relation finder for debugging (tracking down synonymous URIs)
<mscottm> EricP: the relation finder isn't meant for that right?
AnjaJentzsch: we need to work on link quality too
AnjaJentzsch: relation finder is more for browsing and data end-use, less for debugging ... considered a faceted browser approach to browse by concepts and properties ... iTrip gave us visibility to browser creators ... have been helping them attach to our data
mscottm: got interesting animals and geography query from my four year-old
mscottm: couldn't work out how to find the names of things
mscottm: so couldn't figure out if raccoons live in india -- got stuck on finding both identifiers
AnjaJentzsch: posted demos for relation finder, e.g. links between Einstein and Geoddle
AnjaJentzsch: the dbpedia link could help you find the identifiers for raccoons and india
mscottm: found too many links
AnjaJentzsch: i entered "raccoon" and selected the animal concept ... we don't the area the animal lives in ... relation finder can help you find the genes associated with the most common diseases COI?
<mscottm> ericP (COI): Meeting 2 weeks ago to prepare for the AMIA presentation and demo
<mscottm> .. that will happen in San Francisco, California a few weeks after the HCLS F2F
<mscottm> ..Helen and I are looking into the security extensions to the COI demo so that
<mscottm> ..you can see the credentials in the query that allow access
<mscottm> ..working with Fati Turkman (sp?) to create SPARQL constructs for use cases that are met by XACML, which is a standard that is used by interchange between (some?) hospitals.
<mscottm> ..Use case is: some doctor in a clinic asks the database for a patient record. terminology
ericP: is AlanHobbs working on kaiser permanente's SNOMED-OWL?
<AnjaJentzsch> have to leave now, bye
AlanHobbs: no, that's Peter Hendler
<mscottm> Thanks, Anja!
ericP: would like to meet with him Translational Medicine
Bosse: in intensive work to work on a paper ... TMO focusing on selecting use cases ... plan to use the datasets from LODD
<michel> 1. terminology is cross mapped to NCBO bioportal
<michel> 2. terminology is being concretized in OWL file, mapping to BFO / RO, IAO ontology
<michel> 3. marking up use cases with ontology terms
<michel> 4. still need to identify data to cross the use cases
<michel> 5. generate simple application to ask questions
ericP: looked at a PDF poster for TMO ... where is the source ontology?
ElgarPichler: have whittled down to the minimal subset
<michel> current file: http://esw.w3.org/topic/HCLSIG/PharmaOntology/Ontology?action=AttachFile&do=get&target=tmo-20090924.owl
ElgarPichler: mostly refing now ... would like to compare our ontology to yours ... expect an update to our ontology today or tomorrow scientific discourse
mscottm: propose to publish the current drafts of the scientific discourse
APPROVED: publish the current drafts of the scientific discourse documents as an HCLS IG note
<michel> q: how will this work with IAO group?
<michel> larry hunter, etc
mscottm: IAO = information artifact ontology ... larry hunter now participating in scientific discourse calls ... expect he will represent the info artifacts use cases (e.g. i found this info in this document)
mscottm: larry said he will assign folks for collaboration F2F
mscottm: mark musen (NCBO, Stanford) and NextBio will give talks
<matthias_samwald> (need to leave, bye!)
<mscottm> thanks matthias!
mscottm: expect a use case discussion with SPARQL re: Update, property paths, query federation
ericP: query federation slightly different from recent BioRDF paper; encompases that use case
mscottm: when we arrive at the f2f, we'd like to present use cases ... so please gather said use cases ... expect to talk/learn about OWL [QL] ... a provenence IG is starting ... use case: examining evidience for a set of triples ... please feed us more use cases ... also, who's going?
<michel> cannot go
Julia_Kozlovsky: negotiating, but unlikely
<michel> would like to participate by phone, when possible (i teach on both those days)
<mscottm> Egon - can you go to the F2F?
<bbalsa> Not likly to be able to participate.
<egonw> no, not oversees
<egonw> will make it to amsterdam, though skos testimonial
<egonw> hoping there will be a F2F there
mscottm: terminology task for published a skos testimonial
mscottm: expect more at applications in health care and life sciences workshop in AMS
<egonw> I look very much forward to discussing RDF in relation to Bioclipse, for visualization at a molecular level...
mscottm: last year inlcudes lots of interesting use cases
ADJOURNED Summary of Action Items [End of minutes] Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log) $Date: 2009/10/01 20:14:38 $