InformaticsMaryland 2008, JHU's Shady Grove campus

• breakfast, talked with Declan Murphy of Mantaro Networks

Pradeep Ganguly (Mont. Cty. Dept. Economic Development, see smartmontgomery.com)

• shady grove technology campus a 35-year investment
• expected expansion of JHU campus

first session

• Steve Salzberg (UMCP)
  • influenza genome sequencing and analysis
• Ivor Knight (Canon US Life Sciences)
  • signatures to identify pathogens, using insignia pipeline
  • francisella spp
  • did he say something about microfluidics?
• Jacques Reifman (US Army med research)
  • several different pipelines - signatures, function assignment, structure prediction, docking-screening
• Ken Lohman (Midwest Res Inst, MRI; the one the staffs the renewables research center in Colorado)
  • single-molecule mega-sequencing technologies in the works
  • genomic sequencing still driving development
  • upcoming bioinfo needs are in genome annotation, metagenomics, and systems biology
• ? (Sanaria), substituting for Hoffman
• Mihai Pop (UMCP) gut biome
  • shotgun reads of gut biomes of two individuals
  • about 600 different microbial species

break

second session

• Christopher Sprangel (medimmune, bought by AstraZeneca) information management
• Jian Wang (BioFortis) translational research, bench to bedside, personalized medicine
  • labmatrix utility: from data to decision (triangle with decisions at top, data at bottom)
  • currently this gap is filled by Excel (used everywhere), Access, Filemaker, ad hoc systems
  • integration problem: bring together data from individual streams to make decisions
    • demographics
    • expression
    • proteomics
    • clinical data
    • etc.
  • each individual stream is handled well, problem is query-retrieval accross types of data
  • usually done with Excel, his alternative is the labmatrix utility (not well described)
• Randy Ribaudo (Human Workflows, LLC)
  • various sources of expt'l and clinical data
  • same problem as before, problem of integration
  • integration needed to generate new knowledge
  • silos to solutions: another triangle with
    • "siloed" technology at bottom
    • integration layer
    • human interaction (solution has to fit how humans work)
  • case study - prostate cancer
    • ends up with Labmatrix utility of BioFortis (previous speaker)
    • found 39 proteins associated with cancer, 1 protein associated with invasive prostate cancer
• Greg Bertenshaw (Correlogic)
  • biomarker discovery and development company in rockville
  • premise of company is that individual biomarkers are poor predictors
  • "KDE" system-- clustering, centroids, genetic algorithm
• Anna Glodek (Avalon pharma, was at TIGR, met while talking with Steve)
  • compound profiling: gene expression profiling applied to drug discovery
  • example of cancer cell line: uses affymetrix genechip technology
  • challenges
    • data integration: internal and external, refs, pathways, etc
    • methods: developing custom statistical methods
    • tools: choosing best software tool
    • process: implementing a process that is meaningful intuitive and robust
• Mike Fannon (Michael Fannon Associates, Inc) generalizations on state of industry
  • from genes to drugs (you can't get there from here)
  • every new biologic means you are developing a new supply-manufacturing chain
  • clinical portfolio management; clinical regulations push you toward "siloed" data
  • challenges in acquiring bio knowledge
    • complexity of biol systems
    • quantity of info
    • variety of info types
    • semantic problems - different names (somatotropin = growth hormone)
    • IP management
    • biological research culture
  • IT framework for bio research
  • funny derivation to account for executive salaries in the light of executive ignorance
    • power = work/time;
    • substitute: time is money; knowledge is power
    • solving for money: money = work/knowledge
    • conclusion: as knowledge goes to zero, money goes to infinity no matter how much work is being done