Moore Notes 10 9 13

Group Call

• Participants: Katie, Jonathan, Guillaume, Ladan, Sarah, Stephen, Josh, Patrick

• IMG cron job to update databases
  • No API for getting IMG genomes downloaded
  • Jonathan will contact Nikos

• Guillaume: PICRUSTs
  • Successful result running it on one sample
  • Tom will run the metagenomes associated with the same samples using the same annotation tool
    • Compare to assess how well PICRUSTs maps 16S to proteins

• Patrick: MetaCyc
  • Formatting MetaCyc so that it can be used with ShotMap to annotate metagenomes
  • Simulations to see how it performs
    • Families are very partitioned, so worried about misclassification
    • Has good precision with sequences in MetaCyc
    • Will next look at genomes from BioCyc that aren't in MetaCyc for external validation
  • Jonathan: need to check the phylogenetic diversity of the genomes
  • Stephen: don't need high coverage for precision-recall, but maybe for other performance measures
  • Stephen: check out minpath for prediction of pathways from enzymes
    • Patrick: wants to start with smallest pathway unit (1-1 or 1-small number)

• Josh: working on Jack Gilbert's Gulf of Mexico sediment study
  • Rasters of sea floor metadata