BioRDF Teleconference -- 8 Oct 2007

<scribe> Scribe: DBooth

HCLS F2F (Agenda, Location, Registration, etc.) - EricP

Susie: F2F is still goign ahead on the afternoon of Nov 8, and all day Nov 9. Agenda on wiki is still subject to change.

<Susie> F2F Agenda http://esw.w3.org/topic/HCLS/F2F/

Susie: But only in the times of when items will be covered.

<jar286> can we update an already submitted reg form?

<alanr> yes

Susie: To register, you need to fill in the W3C form about the Tech Plenary and the AC meeting. There is also a special guest status for attending if you are not a member of the group. Send me email if you need that.

SfN Poster Update - Don

Don: I have a draft (rough) and it's unfortunately big.

<Don_> http://www.brainstage.com/SfN2007Poster_Draft.pdf

Don: It's about 2.5MB PDF.

<alanr> yup 20M

Don: I'll put abstract out later. We've written about the demo search and wanted to put that into it. Right now I pull a couple of pieces from other sources out there.

<Don_> http://www.brainstage.com/CreateTriples.jpg

Don: There's one unique piece I added in a jpg, trying to show how we bring all this together. I'd like feedback.
... I'm hoping to add searches other than what we wrote about, beyond that.

Alan: Looks nice. Creating medline triples dataflow looks good. I'd include the definitions of dublin core and science common ones. And in the ontology to database maps, it could be labeled a bit easier. The DB is an XML record. Instead of saying a URI literal, say something like node or tag.
... Maybe a separate column like the tag or the path (to the XML doc) and the actual value there as the literal, but call it "value" not literal.
... And a separate table that is value in the XML and value in the RDF. In the case of Title and Date they're literals, but in the case of the pmid it's a URI.
... And for dublin core and science commons make it a link to the ontology.

<AdrianP> +1 for Susie

Susie: Simplify if possible. Perhaps the triples. If there were a gap between the triples it might make things clearer.

jonathan: You can't write qnames. Just create a prefix for pmid.

Susie: Even an asterisk at the bottom could say the prefix.

Jonathan: Cheat saying that a prefix is purl.org/science etc.

Don: Mostly focused on what the scientist wants to know.

<alanr> how big is the poster going to be? 4x8?

DBooth: What do we wish to assume about what the viewer knows about RDF?

<jar286> @prefix pm: <http://purl.org/science/article/pmid/>

<jar286> pm:10540283 is a curie but not a qname. ok in sparql but not (yet) in turtle.

<alanr> are other people getting breakup in the audio?

<AdrianP> yes

<jar286> jus some noise, no breakup yet

DBooth: Perhaps annotate the RDF and SPARQL jargon with plain language re-statement of what it is doing.

Don: Scientists will want to know why this is advantageous to use.
... A friend said something useful would be to do more targeted searches, to reduce 8000 papers. The only place the info appears is in the methods.
... Right now she has to read the methods of 8000 papers. Can we do that now?

Jonathan: Not now.

Alan: This will either be through text mining or explicit annotation. Text mining is complicated by the lack of common lexicon for a-beta.

Jonathan: If the GO curators have done their job well, could we go backwards?

<AdrianP> GoPubMed does text mining on abstracts of papers from PubMed

Alan: I havent' seen specific forms at that level ever annotated.

Don: That would be a herculean task to do by hand curation.

<AdrianP> http://www.gopubmed.org/

Jonathan: But they're already doing hand curation.

Alan: You could go backward through the evidence of who links to that paper. But if you can record the specifics of this case it would be helpful. State the problem of what you did and couldn't do.

<scribe> ACTION: Don to write up problem of looking for specific a-beta papers [recorded in http://www.w3.org/2007/10/08-BioRDF-minutes.html#action01]

Susie: Would be nice to set the context more -- e.g., hard now to query data across data sets. We'
... We've integrated a dozen datasets and can now query across them.

Don: I'll take these comments and show it again next week.
... Everything I've done so far is on the web site, in terms of setting up the database.

Conferences for BioRDF (ISWC, WWW) - Susie

<Susie> http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Conference_Possibilities

Susie: This is a starting place -- not comprehensive -- for asking if we're doing something for this meeting or wish to.
... There is an EMEA annual meeting. Some of these are intended as placeholders for potential participation next year.
... BOF mtg
... Organizers of ISWC are providing space for BOF mtg. I've sent email to about 100 people in asia, and lists. Please reach out to people.

<alanr> SEMANTIC SCIENTIFIC KNOWLEDGE INTEGRATION

Susie: We've hit China in a big way. Taiwan, Malasia, Japan, Thailand, etc.

<alanr> AAAI Spring Symposium Series Workshop

<alanr> MARCH 26-28, 2008, STANFORD UNIVERSITY., CALIFORNIA

Susie: There's a call-for-participation aspect. We'll likely only have a room for 1-1.5 hours, so likely there would only be time for 5 minutes of people saying what they are doing.
... Also other mtgs coming up.

<alanr> Integrative Bioinformatics Conference 2008:

<alanr> http://meetings.ipk-gatersleben.de/ib08

Susie: We've missed the deadlines for some, but might be worth asking about participating.
... We submitted something to WWW2008 last monday.
... There was discussion about whether we should submit a tutorial.

Alan: This came up on the coordination group call. My current thought is it might make sense to hook up with SPARQL tutorial.

<jar286> alan, your voice quality is fair (ie not very good) fyi

Susie: Chairs of W3C SemWeb groups participate in a coordination group call. We mainly discussed tutorials. Ivan Herman said it would be nice if there were one around SPARQL, GRDDL, etc. We suggested a LifeSci tutorial around the knowledge base.

Alan: Or a SPARQL tutorial with HCLS as the vehicle.

Susie: Tutorial proposal needs to be submitted way in advance, then slides about a month in advance, but you can still tweak it in advance up until your presentation.

DBooth: Sounds like a great idea, because it could act as a good model for getting people from HCLS into this.

Adrian: We have a running system for GO pubmed.

Alan: Which conf is best? Also there may be a language issue that makes it hard to get difficult material across.

Jonathan: They may have simultaneous translation.

Alan: Opportunity cost: which forum would be best?

Susie: I think it would be very valuable to put this kind of tutorial together. This particular meeting isn't targeted toward lifesci people. But if we do get our workshop accepted it may drive people to the tutorial.

Alan: Deadline for WWW2008 is Nov 1.

<scribe> ACTION: Susie, Jonathan and Alan to work on tutorial proposal for WWW2008 [recorded in http://www.w3.org/2007/10/08-BioRDF-minutes.html#action02]

(Susie discusses other venues listed in http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Conference_Possibilities )

Susie: Want to highlight Bio-IT World conference: http://www.bio-itworldcongress.com/index.asp . I'm one of the organizers. Call for papers just opened.

<scribe> ACTION: Alan to prepare submission for Bio-IT conference [recorded in http://www.w3.org/2007/10/08-BioRDF-minutes.html#action03]

URI note - Jonathan

Jonathan: Please read the draft.

<Susie> URI Note URL http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Tasks/URI_Best_Practices/Recommendations/SputnikDraft

Jonathan: I'll try to respond to comments within 24 hours. Preferably not all commenting at once!

DBooth: I'll look at it today.

Jonathan: The page itself says how to make comments (separately).
... If others could select a date for review it would help, so they're not all at once.

Susie: I'll look at it on Oct 16.

Alan: Regarding BIO-IT world, I just got email asking if I would organize a half day tutorial.

Jonathan: Science commons has hired someone to work on the triple store. It should get quite a bit better in a few months.

Knowledgebase extensions - Alan

Alan: Got Don's installation and the installatino at Derry, so that's exciting.
... It would be great to work with you (AdrianP) to get the data into triple store.
... Who is responsible for generating that RDF? How to go forward?

Adrian: I'd have to find the right person. We're also working on a rule engine. It uses an Enterprise Service Bus. You can HTTP GET/POST and SOAP and SPARQL.

Alan: Can you give us a presentation on Prova?

<scribe> ACTION: Adrian to prepare presentation on Prova for next week (Oct 15). [recorded in http://www.w3.org/2007/10/08-BioRDF-minutes.html#action04]

<alanr> Prova

Documentation of SenseLab conversion - Kei

<Susie> ADNI http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Demo_Thoughts/ADNI

Susie: Data access policies won't allow us to do much with it. Also looked at ADNI dataset. Alzheimmer's disease neural imaging.
... Samples are being collected from anolytes in blood and urine etc.
... And Lots of brain images being done. Structural MRIs, etc.

<jar286> bye folks, wish i could stay, but gotta go.

<alanr> bye

Susie: People are trying to link the data together.
... There are data access policies we'd need to adhere to, but this dataset is worth exploring further.
... Please look at it so we can discuss next week.

ADJOURNED

<alanr> One more action - wait!

<alanr> oops never mind - I see you've got it

<alanr> bye

- DRAFT -

BioRDF Teleconference

8 Oct 2007

Attendees

Contents