Moore Notes 9 2 09

Group Call

• INFERNAL/OTU analysis (Tom will post update on wiki)
  • have 16S reads from GOS
  • how to align?
    • INFERNAL aligns reads to probabilistic model
    • has an option for shotgun data (--sub) that does best local alignment if read is short
    • what profile to use? makes a big difference
      • native (just 5 sequences)
      • from your alignment
      • RDP hand-curated bacterial and archaeal alignments
    • Sam: how different from hmmalign (from AMPHORA)?
      • hmmalign uses hmm
      • INFERNAL uses scfg (better for RNA)
    • even short (<100bp) reads align well
    • to do:
      • like STAP, classify first (by domain) - need euk 18S profile, have 100 seqs from STAP
      • size threshold may not be needed
      • post-processing (OK to drop reads since they aligned independently): big gaps, amount aligned to profile (big indels)
  • next: phylogenetic analysis
    • Steve: run through raxml, NJ (?)
    • MOTHUR

• Tom: looking for a dataset of all sequenced microbial genomes to do gene family analyses on genes that are not in AMPHORA
  • Srijak: ComboDb
    • 6 months old
    • coding regions only: what protein, NCBI accession
    • Martin's personal version of db may have the flat files
  • Morgan: microdb
    • genome project
    • chromosomes
    • gene level: protein and DNA sequences
    • updated monthly with cron job

• Sam: simulation update
  • posting data and code on edhar
    • 20 runs for RPOB, doing a few other genes (SMPB)
    • what else do people need?
  • problem with private html directory (Morgan too)
    • Russell will work on it

• version control on edhar
  • git and svn are installed (can run over ssh)
  • to share with people who don't have edhar accounts, use public html for now
  • have group iseem directory - Russell will put an svn directory there with code and data subdirectories

• Authorship discussion
  • Jessica did a literature review
  • She and Katie have discussed on PI call last week, plus some email with Jonathan
  • issues:
    • who is an author?
    • what is author order?
    • other issues: e.g. contributions
  • idea: on next call, discuss 3-4 papers together to set the plan, e.g.
    • simulation paper
    • SAR data paper
    • depth gradient paper
  • Sam: need a model for how things change and evolve during the project