Moore Notes 8 27 14

Group Call

• Participants: Katie, Jonathan, Tom, Stacia, Josh, Dongying

• Illumina meeting Thursday

• Dongying: How to automatically update PhyEco markers slides
  • Universal and even within a taxonomic group
  • How to pick a bit score cutoff for determining new members for each marker family?
    • Need an automated method to find drop in bit score distribution
    • Then optimize universality*evenness*single-copy-score (u*e*s)
      • Katie/Tom: Why not optimize this directly?
  • Sequences that don't get added by these criteria may belong (short branch in gene tree)
    • Automated procedure for building alignment and tree to check for membership
    • Use HMM consensus (hmmemit) to root tree, compute root to tip distance
    • Automated cutoff on root-tip depth to identify outliers (in terms of standard deviations from the mean for the family)
      • Katie: Use false positives and false negatives (or u*e*s) to pick a cutoff
  • Compared u and e of new vs. old families

• What is a good statistic for measuring the "size" of the tree for a gene family?
  • root-tip distance
  • number of leaves
  • root-tip/leaves
  • sum of branches (normalized?)
  • density of tree calculation that Dongying developed

• Upcoming calls:
  • Sept 3: Shotmap
  • Sept 10: Sarah?
  • Sept 17: No call