Moore Notes 9 8 10

Group Call

• Quarterly report due Oct 8
  • JE should make a template

• Projects using GOS data
  • Steve
    • Assign reads to AMPHORA gene families with HMMR/AMPHORA
    • Build phylogenies for each gene family
    • Phylogenetic diversity analyses
      • alpha: within sample
      • beta: between samples
      • connect to environmental variables
    • Main results: geographic region (large scale) and habitat (e.g., coastal vs. open ocean) explain most of the variability in beta-diversity. Within region, salinity and some other ecosystem variables are important predictors of diversity.
      • Josh: Contradicts macro-organism results from Tittensor et al, but habitat finding is consistent with Jed Furman's work
      • Could be confounding based on changes in methods (e.g., filter size, sonication vs. restriction enzymes for library preparation) across regions
      • Katie: try adding variables for trip or filter size to ANOVAs
  • Sam's simulations
    • might shed light on whether or not beta diversity estimation is affected by problems with trees (errors might cancel out)
    • will discuss community composition off-line: can sample from a tree or can change abundances
  • Taxonomic diversity (Josh and James with Tom's OTUs)
    • methods note: PHYLOTU OTU abundance rather than PD
    • James: very little beta diversity signal
      • could compare to salinity and temperature
      • might be able to account for detection biases (e.g., from Jenna's project)
    • Josh
      • Niche modeling for GOS data, might be an example in paper about estimators
      • Environmental niche modeling used GOS as a preliminary analysis, might include with other 16S data
    • Tom could use GOS for calculating total species richness (but mostly working with full 16S data set)
  • Protein family diversity
    • Tom: taxonomy vs. gene family phylogenetic signal - might use GOS
    • Josh: niche modeling of Operational protein families
    • Morgan
      • diversity measurements on protein families (presence/absence, maybe abundance)
      • correlations with environmental meta-data
      • Steve: very parallel analyses to the PD work
      • compare to taxonomic signal

• Fort Lauderdale regulations on genome sequence data restrictions
  • http://www.genome.gov/10506537
  • Katie: Easy to generate a genome now, but still hard to analyze so consortia continue to exist
  • People are sitting on genome data and not publishing the genome paper (e.g., mammals)
  • People are often hesitant to share data they generate (e.g., VIROME, HMP)
  • JE blog / PLoS Biology editorial
    • researching it, e.g., for HMP sequences are getting released pretty quickly
    • meta-data is taking longer: hard to annotate, privacy issues (some human DNA even after filtering)
    • Rob Knight: open microbiome project
    • HMP DAC site (Owen White) vs. Genbank
    • mammal genomes (Aaron heard about problem too)