HCLSIG/PharmaOntology/Meetings/2011-01-06 Conference Call/minutes
On IRC I see michel, Bob, mscottm, matthias_samwald, bbalsa, Joanne, Zakim, chimezie, Bin, ericP
Bob scribenick Bob
Michel: EricP points out that COI shows how clients can use TMO, but we don't get feedback
... want to close that circle. Want to put action items on our plate when feedback
EricP: There's not that much happening yet. Looking for grant money
... COI Clinical Observation and Interoperability, original task was to take EHRs and use RDF representation
wiki page: http://hcls.deri.org/coi/demo/
... one step up, showing how easy RDF would be to connect
... deliverable was a patient recruitment demo. Anonymized
... protocol can be partially expressed as sparql, can translate whether HL7, RIM, etc
... Other COI work has been focussed on access to patient records (liquidity of data)
... figure out how to leave data in situ, use RDF for communication
I2B2 is one of the NIH's National Centers for BioComputing
... Indivo is focussed on one patient. xml, grddl to RDF
Michel: What can we do to further these COI-activities
EricP: Let them use it for a while
Michel: Relationship w. Indivo. SNPs are very interesting in our records.
EricP: Most of the mappings are simple. But we need to add structure more; this is harder.
... propose XML structure back to Indivo would be good
... make sure that both Indivo and I2b2 record these more complex data
Proposal: To augment indivo/i2b2 with genomic / variant data and set the corresponding RDF representation that maps to TMO
+1's are coming in.
chimezie I'm concerned about resistance to suggesting people extend their patient data syntax rather than starting with demonstrating value in using the formats they have
Scott: Have learned that there already is practice for SNP info for just the target genes
Joanne: There has been work on sequence ontology; want to use currently existing standards or practices if possible
Michel: Not really SNP standards yet, just SNP vocabularies
Joanne: Not want to re-invent wheel
Michel: Joanne action, find out SNP/genomic? practices
Michel: Care about more than SNPs; want genetic and genomic structure too
Michel: Chime's question
Chime: i2b2 relational format, we might get some push-back if we suggest changes
... want a minimal solution to feed back to i2b2
Michel: Once you have it there, then we show the easy mapping
... we should focus on what would it look like in RDF
EricP dicussing context for blood pressure measurement
EricP: SNP info had a lot of structure, contexts are all the same in gene tests
... so SNP context; we need not extend
Chime: Do we have info about why Indivo is the way it is?
EricP: Have some contact w. Indivo people
Michel: Good point: Can be get a MOU with Indivo? Perhaps a formal relationship?
... coming from a visibility perspective.
Bob: ++1
EricP: Might get funding for i2b2 DB mapping stuff, Children's
Joanne: Support makes things go faster
Michel: Part of objective is to be more specific about EHR, plan for more visibility
Michel: New areas for our group: Animals models
... Christi, Matthias and EricP on animal models
... What exactly do we want to do? What resources do we need?
ericP animal models like ?
Christi: Initially talked about which species would be most appropos
... use animal models for safety and adverse events prediciton
ChristiDenney Found this - interesting.... http://www.lamhdi.org/about
ChristiDenney LAMDHI = Link Animal Models to Human DIsease
ChristiDenney LAMDHI - We use OMIM to link human diseases with specific human genes.
EricP is our go-to-guy.
MGD RGD are the two data bases
Mouse Genome and Rat Genome Databases
Related to Toxicology DBs
Michel: Tox DBs plus comparative genomics DBs
... look at gene neighborhoods and regulatory networks
Scott: Seems like animal models adds another level of work
... relate animal model back to human research
Michel: dbSNP contains animal + human
... some SNPs are homologous w. animal models
Michel: This is for looking back into the research domain.
... looking for animal trial that preceeds clinical trial
Scott: Salvaging research that might have gone unnoticed
Michel: Or pointing out a higher priority for some MOA etc
... Christi can continue exploring, how much data and what do they look like
Michel: Joanne re paper
Joanne: Got a response back that there will be a response later.
... CHALS targeted at pharma. EricP will present a tutorial.
... (correction here?) RPI will be doing tutorial
... as soon as we have a tiny demo, using Michel's sparql endpoint.
... Also submitted abstract to AMIA in San Francisco
EricP: At CSHALS I'll be talking about EHRs
ChristiDenney NINDS/Parkinson's - http://www.ninds.nih.gov/research/parkinsonsweb/amr/animal_model.htm
Michel: Off to a good start for the new year!