See also: IRC log
scribenick bobP
<matthias_samwald> https://docs.google.com/document/d/1bsJ3YWuH1k3gdXtSN2PXfmUIavtR-hDXhHMn0bKB5pk/edit?hl=en_US
Michel: Write short special report, then consider the new charter
:)
Michel: Scott concerned w timing, Matthias requirements
Matthias: Agreed on using this
kind of data type Special report 1000-3000 words
... google doc started
<mscottm> About how many pages is that?
Matthias: from the viewpoint of
normal research, personalized med, we can say that semweb is
niche area
... requirement for future medicine, article should give
overview of problems that motivate
... also non-experts need a good overview of what we are
suggesting
... brief enough to focus on ideas and mission
... Outline in google doc is summary, not really an
abstract
... one section is required for future perspectives, where most
of text should go
... looks out 5-10 years, how semantics, text-mining etc will
play
<ericP> ciao
Matthias: we can agree on high
level statement, then go on discussing how pharma would be
involved
... and integrated into clinical decision support
... can also present ontologies, say how we expect pieces of
puzzle to fit in infrastructure
... over next two weeks
Michel: Good, like the high-levelness and tone
<bobf> Is there a deadline for the submission?
Michel: interesting to present
what semweb can do
... pasted in resources at heart of integration
... should also have a figure
... don't know the value of including everything
Matthias: Have already reduced table somewhat
Michel: Figure: variation,
response, pop genetics, diseases and treatment
... table has raw data, as terms and onts to leverage or
describe
... maybe illustrate the RDF-like character, or a concept map
about relationships
<ericP> i don't feel qualified to answer
Scott: Solves the hesitation in
the model
... position paper, put it out there, then continue in
research
... should mention gene-variant concept in this paper
anyway
Michel: Starts w future
perspective, can talk re resources and how can be
exploited
... Need to come up w model on how to make contributions
... suggest one or two people to write first draft
... message needs to be highly focused
... Matthias and I to write first draft?
... by next week - Yes.
... next Thu will have coherent draft
<mscottm> Scott was referring specifically to the 'gene variant pipeline' concept from Matthias in the earlier draft as nice to include in the position paper.
d'accord
Michel: Talk re research
heading.
... Summary of competency questions
... have gone through competency questions
Matthias: Have a table.
<michel> https://docs.google.com/document/d/1I9xyVKhO9wG7My2fWu9S-wMwWZxG6PkRc98kOxUqH-o/edit?hl=en_US
Michel: Need to finish relevant
datasets
... also need application and data services need to be filled
out
Matthias: We could try to use
extended version of TMO
... owl reasoning on top to classify a particular patient into
a stratification(?)
Michel: Task, create nec and suff conditions for classification
Matthias: Should agree on
extension of TMO
... while ago I worked on this; should have a separte call (to
nail this down)
Michel: Data representation
Matthias: TMO is realist-based,
another is generic RDF-izaiton, not nec the same
... should do both in parallel, but with connections
... generic RDF, then try to map to TMO that can be condensed
for reasoning
Scott: Going back to original
implementation, how to represent SNPs etc
... more viable form of representation,..
... need a way to represent SNPs in relation to patients
... relation drug/snp/dosage/phenotype
Michel: Overarching diagram,
maybe continue developing
... how data are connected should be explicit for us
... need to review that diagram
... Had started a diagram w types and relations between
them
... now we understand how things are connected and hence how to
construct querie
Matthias: Diagram for shot paper needs high-level concepts
Michel: Linked data ontology project needs specific enties, etc
Matthias: Diagram + Protege, two different tracks?
Scott: top-down vs bottom-up
Michel: Ontology is
implementation, to think about info. It's the artifact.
... diagram is conceptual, how things ought to be.
... also shows how to construct queries
Scott: Diagram shows info requirements; then move to ontology
Michel: If we agree on diagram, then we move to ontology
ericP: Interface, pattern 1 -> pattern 2, so picture shows pattern 2?
Michel: Yes.
<michel> diagram: https://docs.google.com/drawings/d/1OaqlmiTbIAkgHM6kCV-4xbRddHJiN1YZkCWdNHvwiGI/edit?authkey=CIT4rcwM&hl=en_US&authkey=CIT4rcwM
Scott: Will be working ont ->
sparql; some things better handled in owl?
... patient instances are in KB
<iker> my regrets but i must leave
Scott: TMKB, using next to
dbSNP
... Any possibility of getting clinical data or sample SNP?
Matthias: SNP datasets are out
there, also genomes
... need work to extract. In public domain
Scott: No machine learning. Need realistic forms of data
Matthias: Cool to create view of patients w 500K data points!
Scott: (scared?) but could
work!
... SNPs that are relevant to treatment, assume identified
early
Matthias: Can have combinations of SNPs that have special meaning
Scott: Machine learning in addition to these other connections?
Michel: Focus on table of competency questions
<mscottm> Scott was seeking assurance that we won't be attempting machine learning at this stage. :)
yes :)
Michel: Add competency questions, one by one.
Scott: Synch w dbSNP. Next level might be to get sql dump on disk
ericP: Iker and I both worked on
this.
... scripts complicated. Saw ftp site etc, easy to get a little
bit, but dbSNP is huge
... wget() does not do the whole thing
... Oxford has scripts, which included sources
Scott: Oxford has given a way to
load into mySQL, brought Iker to them
... ideally us SWObjects w database connector
... once working, we could talk sparql endpoint for dbSNP
ericP: Representative sample
would be awesome
... now foreign keys declared
bobf: Can help guide though the schema
ericP: For one use case, we can work through which columns we need
Matthias: dbSNP does not answer
specific clincial questions
... know different SNPs available, different alleles and
percentages in hapmap
... but don't need to know all the references etc
<mscottm> +1
bobf: Show semweb capabilities or
create new knowledge?
... pgx people are wary of inferring too strongly!
ericP: Goal is to say, using
semweb will make your job easier.
... maybe recapitulate and existing paper
bobf: List of resources here is a
really good start
... here are areas of discrepancies, ..., and here is how we
would solve this
Michel: dbSNP use case to connect
data. SNP -> other resources connection is important
... relations sequences, gene products, pop variation etc. Need
use cases to drive this.
Scott: By making available in
RDF, can integrate across datasets. SNP db different from drug
db
... usable in clinical research, pharma. All in one software
landscape.
Michel: Matthias and I on first draft. Please add a use case to the table.
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) No ScribeNick specified. Guessing ScribeNick: bobP Inferring Scribes: bobP WARNING: No "Topic:" lines found. WARNING: No "Present: ... " found! Possibly Present: Don_Doherty Donald_Doherty Matthias Michel Scott aaee adrien bobf diagram epichler ericP https iker matthias_samwald mscottm You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy 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: 15 Sep 2011 Guessing minutes URL: http://www.w3.org/2011/09/15-hcls2-minutes.html People with action items: WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report[End of scribe.perl diagnostic output]