Moore Notes 5 11 15

Group Call

• Participants: Katie, Tom, Jonathan, Sarah, Patrick, Dongying, Stephen

• GBMF:
  • Software release report:
    • Last June
    • Post new things to wiki
  • Progress report
  • No-cost extension

• Shotmap revisions
  • Expanded IBD analysis with population of additional samples
    • CD separates out from UC and healthy in PCA plot
    • No longer see lower functional richness
    • Patrick will repeat statistical tests
  • Tom working on software items
  • Stephen is handling simulation items

• Tom: clade distributions project
  • Phylogeny aware microbial distributions
  • Motivation:
    • Sparsity (correlations maybe more meaningful at less granular level)
    • Multiple testing correction (fewer tests if use ancestral vs. leaf nodes)
    • Properties of clades (versus OTUs)
    • Identifying core taxa
    • Co-diversification of microbiome with host/environment
    • Hypotheses about interacting clades
  • Related work:
    • Methods paper: http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1004186
    • JE: What about plant/animal literature? Eco-phylogenetics, intra-specific phylogeography
  • Proof-of-principal study with Muegge et al (2011) data on six non-human primates
    • 11 clades common to all samples (one within Prevotella)
    • Clades co-vary across samples
  • Tree building and traversal is slow and error prone
    • Traverse reference tree
    • Place reads (slow)
      • Iterate over phyla?
      • Classify reads into reference OTUs (green genes tree)
      • Need to validate accuracy
    • Build clade abundance matrix from edge to clade matrix
      • KP: Use pplacer probabilities rather than hard assignments
      • Patrick: non-independence of tests
  • JE: Apply to marker genes in metagenomes, not just 16S, for more novelty and resolution