Moore Notes 7 14 10

PHYLOTU group call: Pollard lab, James, Steve

• Name: PhylOTU? PhyloTU? PHYLOTU?

• Tom will send submit version before sending in

• Richness underestimation issue
  • PID vs. PD distance matrices (508 full length 16S sequences)
    • Strong correlation
    • PID is bounded by 1, and most are around 0.2
    • PD is not bounded, different skew
    • Normalized PD distance matrix doesn't improve richness ratio - not a linear difference
    • What is going on in the range around the threshold? Try stratifying
  • PID vs. PD trees (fastree or NJ)
    • What matters is the comparison of branch lengths around the threshold, not rest of tree
  • Look at hierarchical clustering trees and ordered distance matrices