Moore Notes 10 23 13

Group Call

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

• Guillaume & Tom: PICRUSTs update
  • Samples processed with PICRUSTs
  • Metagenome annotations will come off cluster today
  • Correlation analysis, other validations (e.g., histogram of disagreement values)
  • Stephen: what do we expect? Tom: depends on phylogenetic distribution of families, degree of vertical inheritance, family abundance
  • Since L4 English channel data is 454 sequencing, not super deep coverage on metagenomics side (i.e., low abundance families won't be well estimated in the metagenomes)
  • Stephen: How would the data be used? e.g., relative abundance vs. presence-absence, total diversity vs. individual families

• Katie will check with JE about contacting Jeroen re: Tara Oceans

• IMG download issues
  • JE contacted Nikos
  • Nikos said he'll look into it

• Stephen: MicrobeCensus (estimating average genome size)
  • Slides
  • Motivation: differences in genome size between samples can bias gene family abundance estimates
  • Used a subset of 30 of Donying's marker genes that are universal and single-copy in bacteria and archaea http://phylosift.wordpress.com/tutorials/scripts-markers/
  • Tom: What about 10 copies in 1 genome vs. 10 copies in 10 genomes? Stephen: Method doesn't distinguish these cases, focuses on average genome size and average genome coverage for the community in each sample to correct for variance in relative abundance across samples that is simply due to average genome size differences
  • Other data sets to analyze
  • Look at effects on relative abundance and identification of biomarkers
  • Genome size and variation in environment (e.g., might need more genes to live in many environments)
  • Tom: Do results differ between bacteria vs. archaea? Stephen: No, just endosymbionts without all the markers can cause problems
  • Need to look at other domains
    • Eukaryotes (i.e., are markers single copy in eukaryotes too?)
    • Viruses/phage may cause some problems
    • Could add these taxa to the simulations and just see if it makes any difference