Moore Notes 3 26 14

Shotmap Working Group Call

• Participants: Katie, Tom, Stephen, Stacia, Patrick, Guillaume

• R21 microbiome temporal variation grant reviews were positive

• Summary of project: slides
  • Collaborators
  • Goals
  • Existing solutions to functional analysis of metagenomes
    • Shotmap advantages: optimized read-length specific parameters, local, distributed, flexible, database organization
  • Statistical simulations
  • Empirical analyses
    • Re-analysis of Gilbert's English Channel data
      • Confirm main findings
      • Goods coverage low in transcriptomes with KOs but not with SFams
      • Differentially abundant families:
        • Transcriptomes differ from genomes
        • Oxidoreductases higher relative abundance in transcriptomes vs. genomes
        • Katie: try overall test and then pairwise
        • Patrick: try different multiple testing procedure or lumping families into groups to test first or modeling differences and looking for outliers
      • Clustering of temporal profiles: updated temporal variation slide
        • Patrick: Log transform, plot as heat map, order columns by covariates
        • Katie: Standardize by IQR
    • MetaHIT data

• To do list and who can help
  • Tom: wrap up L4 analysis
    • Stacia: Can we analyze SEED to directly compare to Gilbert et al?
  • Stacia/Tom: Redo MetaHIT samples run by Patrick and Tim with optimized parameters
  • Stephen: Read-length specific parameters
  • Patrick/Tom: Time, memory, and space requirements (and how much db grows with each analysis), e.g., on HMP or T2D or GlucoseControl
  • Stacia: Software improvements: testing, data analysis tools, MicrobeCensus integration

• Revisit this on April 9 or 16 call