Moore Notes 5 15 13

Group Call

• Participants: Katie, Josh, Tom, Stephen, Dongying

• Annotating SFams
  • Got annotation file from JGI
    • Dongying: file not bigger than previous one
    • Tom: will try using this for now, but Tim has Interpro running on the cluster
      • backup plan, or for next update
  • Stephen: building HMMs for KEGG Orthology groups
    • Lots of sequence length variation within groups
    • For genome annotation, he is filtering out long sequences
    • What do we want to do for annotating SFams?
      • Alternative to HMMs would be to pairwise search
      • E.g., BLAST or RAPSearch vs. ALL KEGG annotated sequences (including very long or short ones)
      • Tim will compare and we can decide from there
  • EggNog and KEGG both have some sequences in more than one ortho group
    • Stephen: may be due to fusion proteins
      • Dropping these for simulations
      • Dongying: dropping might not be a good idea

• Tom talk to Victor Hansen-Smith from Sandy Johnson's lab
  • Amino acid substitutions that alter function in a gene family
  • Idea: use his ancestral sequence distributions to search SFams
  • Another idea: unannotated SFams (not connected in SFam network) to study de novo gene creation
    • Rates, patterns on tree, etc.

• Josh: problems in soil niche modeling
  • Tibetan soil study
    • local data: C:N predictive of diversity, pH not (contrary to studies in other places)
    • raster data: pH is correlated with diversity
  • Check local vs. raster for other variables
  • Compare range of local values in Tibet versus N America (same for rasters)
  • Tom: C:N is important in mycology biomass

• June 17 meeting agenda

• Josh and Stephen headed to ASM in Denver