intro

<kei> scribenick matthias

<matthias_samwald> kei: we have been exchanging e-mails in the last few days, lena has been looking into genelist representation

<kei> http://esw.w3.org/HCLSIG_BioRDF_Subgroup/QueryFederation2

<matthias_samwald> ... i made some edits to that page: first, giving high-level context

<matthias_samwald> ... the more detail the metadata/annotations give, the more likely the user is to find relevant datasets

<matthias_samwald> ... another potential benefit of adding metadata to microarray datasets is that it allows us to better compare experiments.

<matthias_samwald> ... and, of course, another major benefit is integration. microarray results describe a lot of different gene expression patterns, but sample size is often limited. pooling could increase sample size, improving statistical power.

<matthias_samwald> ... often you find a selective gene list in papers based on microarray results. a small number of genes is listed in the paper. they are often exposed as human-readable tables, but not in a machine-readable format.

<matthias_samwald> ... i added all these points to the wiki page.

<matthias_samwald> ... at the bottom of this page i also added several queries.

<matthias_samwald> ... still quite general, but a start.

<matthias_samwald> matthias: what is "fold change"?

<matthias_samwald> kei: a ratio that gives information about up- or downregulation

<matthias_samwald> scott: they use significance analysis to see if significant differential expression happened. fold change is intuitive, but it does not provide whole picture.

<matthias_samwald> lena: when we talked about gene lists, is there any other measure?

<matthias_samwald> scott: the p value.

<matthias_samwald> lena: i think the p value refers to the fold change?

<matthias_samwald> scott: no. fold change refers to the raw signal.

<matthias_samwald> scott: often it is indicative, but after normalisation and statistical analysis, some drastic changes in fold change turn out not to be caused by significant differential expression.

<matthias_samwald> kei: i think this discussion is helpful.

<matthias_samwald> ... we are not the ones who can judge the validity of results, but we can use metadata to expose context. this makes it possible for other researchers to judge validity and find possible improvements.

<matthias_samwald> scott: people would like to know where gene list came from, which algorithms were used etc.

<matthias_samwald> ... knowing that can make data more valuable.

<matthias_samwald> scott: recently i had the chance to talk to chris stockert of MGED. i asked him how to push this forward. he suggested to contact publishers.

<matthias_samwald> ... publishers could ask authors to submit information. going back to unstructured text is not feasible.

<matthias_samwald> ... i also talked to [name] from gene atlas; interested in RDF representation, they are facing similar problems

<matthias_samwald> ... we should ask him to join the call

<matthias_samwald> kei: if we don't go through it ourselves, we will have problems with collaborating with others. trying it ourselves will give us a good understanding.

<matthias_samwald> ... some of us will be looking at spreadsheet and try to come up with RDF representation

<matthias_samwald> lena: the different gene lists are giving us differnt data. there is something missing in list 2.

<matthias_samwald> lena: they are only giving us the signal which is only meaningful in the context of the group and is not comparable

<mscottm> http://neuinfo.org/

<matthias_samwald> scott: the main thing of interest is the list of genes itself. just knowing the lists of genes already enables queries.

<matthias_samwald> ... the NIF project would be happy to support our work

<matthias_samwald> ... they can provide vocabulary about the neuroscience domain

<mscottm> http://bioportal.bioontology.org/

<matthias_samwald> ... BioPortal is also relevant

<matthias_samwald> ... if you go to BioPortal and search for 'entorhinal cortex', you will find it in some of the ontologies.

<matthias_samwald> kei: i also recently talked to maryann and steven larson.

<matthias_samwald> ... there will be a SPARQL endpoint for Neurolex (which is different from NIF ontology)

<matthias_samwald> scott: NCBO will also soon have a SPARQL endpoint.

<mscottm> It has one now: http://sparql.bioontology.org:8080/webui/

<mscottm> Also: http://sparql.bioontology.org/webui/

<matthias_samwald> kei: we do not need to worry too much about the quantitative values at the moment. researchers mainly want to know about 'is it under/overexpressed?'. it is mostly a qualitative query.

<matthias_samwald> lena: in some of the gene lists i cannot see the information about whether they were up- or downregulated

<matthias_samwald> kei: the page usually gives some context.

<matthias_samwald> kei: knowledge about some kind of differential expression could already help.

<matthias_samwald> lena: fold change could give some hints on whether a gene is over- or underexpressed

<matthias_samwald> scott: usually scientist do RT-PCR to see if a gene is over- or underexpressed

<matthias_samwald> kei: of course we do not need to exclude those quantitative values when we have them

<matthias_samwald> kei: satya, do you have any recommendations for providence representation?

<matthias_samwald> satya: we are collaborating with sci-discourse group, have shared OWL file, maybe lena could discuss with them as well.

<matthias_samwald> lena: before we go to sci-discourse, we need to make sure it uses existing terms

<matthias_samwald> satya: we are already using external terms, we make it clear that is undergoes changes.

<mscottm> scribenick: scott

<mscottm> Matthias: Yes, the HCLS KB contains the diseasome dataset and all of the datasets that came out of the LODD work.

<matthias_samwald> matthias: nothing new about Science Commons knowledge base.

<mscottm> ..I haven't heard anything from Neurocommons. A new release is supposed to be on the way.

<mscottm> scribenick: matthias

<mscottm> Bye - I have to go to the Terminology call.

<kei> thanks, scott

<matthias_samwald> satya: KEGG, BioCyc etc. expose pathway data as RDF (BioPAX)

<matthias_samwald> kei: the group can also think about publication opportunities.

<matthias_samwald> kei: i will be away between late june and late july.

<matthias_samwald> kei: jun, maybe you could help out during july.

<matthias_samwald> kei: we should try to keep the momentum going. is it okay for everyone to participate in weekly calls?

<matthias_samwald> scribenick: matthias_samwald

This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found ScribeNick: scott
WARNING: No scribe lines found matching ScribeNick pattern: <scott> ...
Found ScribeNick: matthias
WARNING: No scribe lines found matching ScribeNick pattern: <matthias> ...
Found ScribeNick: matthias_samwald
Inferring Scribes: scott, matthias, matthias_samwald

WARNING: 1 scribe lines found (out of 115 total lines.)
Are you sure you specified a correct ScribeNick?

Scribes: scott, matthias, matthias_samwald
ScribeNicks: scott, matthias, matthias_samwald
Default Present: +1.832.386.aaaa, Kei_Cheung, matthias_samwald, [IPcaller], mscottm, +1.937.775.aabb, +1.937.775.aacc
Present: +1.832.386.aaaa Kei_Cheung matthias_samwald [IPcaller] mscottm +1.937.775.aabb +1.937.775.aacc

WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting


WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 24 May 2010
Guessing minutes URL: http://www.w3.org/2010/05/24-hcls-minutes.html
People with action items: