HCLS Telecon -- 10 Jul 2008

<Huajun> Why can't I see me on the phone?

<Huajun> I amfrom a iphone

<mscottm> having skype problems

susie: we will start a "scribe-list" for the teleconferences.
... lets start with introductions from participants.

<ericP> kanzure is Bryan Bishop

[introductions from don doherty, eric P, huajun chen, kei cheung, matthias samwald]

[introduction from sivaran Arabandi (cleveland clinic)]

<mscottm> sorry - going to try reboot to solve "Problem with recording device" in Skype

<ericP> http://lists.w3.org/Archives/Public/public-semweb-lifesci/2008Jul/0015 says that Bryan Bishop is a semweb hacker, currently working with data from the Allen Institute

[introduction from kristen ??? (microscoft)]

[introduction from bryan bishop]

Marketing Update

susie: we need to think about how to "market" the things we as a group are doing.
... chimezie, scott and I had a discussion with a contact at W3C, the W3C will help us in "marekting", e.g., media outreach

<Susie> http://esw.w3.org/topic/HCLSIG

susie: at the moment we are compiling lists of for the most important venues that we should be at.
... see http://esw.w3.org/topic/HCLSIG under the header 'marketing'
... this will help us target papers etc.

<Susie> http://esw.w3.org/topic/HCLSIG/Conferences

susie: at the moment the list is biased towards my interests

<Susie> http://esw.w3.org/topic/HCLSIG/Media

susie: it is important that we do not only focus on journals, papers etc., but also on other media such as radio, the blogging community etc.
... i put up a list of around 15 journals that might be of interest to us.

<ericP> matthiassamwald: when we have identified the channels, how do we generate the content?

<ericP> ... we had challenges generating a colaborative paper in the last edition of this IG

matthias: but how do we create the content we disemminate? we need a streamlined process for collaborating.

susie: people that work on a specific project should have the lead about writing something about their projects
... other participants only do 'polishing'

<mscottm> thanks eric

kei: what about recruitment of personal (e.g. post docs)?

susie: we should try to get 'key people' in the community on board of the HCLS IG.
... posting notifications about open positions to the mailing list is OK.

<Zakim> ericP, you wanted to endorse putting a message after the q+

Introductions

Ideal Attributes of Projects

susie: projects need to have at least some scientific focus.
... HCLS IG task forces will have separate calls
... and dedicated coordinators.
... we would prefer projects that link to other activities, making it possible to have a common "mission"

scott: we would like to have a applied approach in HCLS

<ericP> +1 to framing as scientific question as early as possible

scott: projects should aim at answering specific questions

kei: can a task force span multiple projects?

susie: generally, i expect that there is one project per task force. the BIORDF task could contain several tasks, though.

Representing Clinicaltrials.gov in RDF

susie: ok, let's discuss some ideas for projects.

<kanzure> Hello.

susie: Representing Clinicaltrials.gov --- users might want to be able to do more expressive queries.

<kanzure> Rapid brainstorm of potential semantic web projects?

susie: but integrating it with other datasets would also be interesting.
... Lilly has a website called http://www.lillytrials.com/
... at the moment, most of the information is inside PDF documents.
... NIH orange book is another dataset in this domain
... or drugbank
... this connects to the work of COI
... we should first do a survey on relevant datasets.

<Zakim> ericP, you wanted to ask if we can get our hands on the backend data for Clinicaltrials.gov

Convert the Diseasome into RDF

<Susie> http://www.nd.edu/~alb/Publication06/145-HumanDisease_PNAS-14My07-Proc/Suppl/

<ericP> note that links are not apparent in the style for the above resource, but they are there

susie: there needs to be a specific scientific question when we start such collaborations.
... there is something called "connectivity map" that looks at gene expression profiles of drug responses

scott: it would be interesting to query the association between different diseases via common proteins and pathways.
... there could be some synergy with other projects, e.g. projects that link proteins to diseases via textmining

<ericP> i'm eager to work on this

<Donald_Doherty> I'd like to help.

kirsten: i know that some data sets exist about breast cancer, for example.

<mscottm> Alz, HD, PD: it's a 'marriage of convenience' :)

eric: it is a question of whether we want to branch out or not.

kirsten: but can we at this point still decide what domains we want to deal with?

HCLS KB Enhancements

<mscottm> http://esw.w3.org/topic/HCLSIG/Project_Ideas/KB

scott: it would be good to demonstrate that we can distribute our knowledge sources over several locations.
... the queries at the moment require a lot of specialised knowledge
... one idea to aid in querying is to offer support based on terms in the knowledge base (google-style lookup)

<kei> I have to sign off ...

<chimezie> bye Kei

scott: we are not limited to neuroscientific data, but it also makes sense to build upon what we already have (which is centered on neuroscience)

SUGGESTION: let's start discussion threads about all these issues on the mailing list.

<cgi-irc> I'm happy to follow up offline on the breast cancer data

<cgi-irc> This is Kristin

<mscottm> ok, great.

<ericP> next call: 24July

@eric, could you please tell RRSagent to create the minutes?

- DRAFT -

HCLS Telecon

10 Jul 2008

Attendees

Contents