Moore Notes 1 19 10

Group Call: Steve, Tom, Morgan, Guillaume, Katie, Sam, Josh

• Reports
  • Quarterly report - does Jonathan need anything from group?
  • Annual report due end of Feb

• How to use conference call time?
  • individual updates (1-2 topics per call)
  • leave room for stuff that comes up
  • project conversations (rather than separate time)

• Tracking system is set up - start using it!

• Protein family clustering of reference data (genomes)
  • All vs. all blast analysis is hard computation (due to families with lots of paralogs)
  • About 1000 genomes in microbeDb, extending to IMG (maybe unpublished GEBA genomes)
    • Guillaume (w/ Morgan) is working through IMG genomes
    • Could extend to eukaryotes
    • Trying to deal with genomes that appear twice (draft and final) and other clean up issues
    • Which sequence identifier to use (IMG vs. genbank etc)?
  • Dongying has done MCL clustering with 100 genomes, all proteins
  • How to add more genomes?
    • Use Dongying's families to search for more copies in other genomes
    • Or start from scratch
  • Plan: build HMMs and trees out of these for downstream analyses
    • Tom is working on pipeline to add reads to Dongying's families
    • Morgan is going to look for sequences that don't hit these families and cluster them
  • What features do people want? Talk to Tom et al.
    • Currently planned features: sequences, annotation, alignments, phylogeny (ref seqs only), scores/metrics, rates of evolution, HMM profiles
    • RNA genes (e.g. rRNA for OTU projects, tRNAs),
      • maybe separate db if just SSU rRNA
      • not always well annotated
    • Separate part of db/objects for specific read data set analyses?
  • Get edhar MySQL logins from Morgan if you want to use it

• OTU group update
  • Tom: summary of pipeline
  • What data (besides GOS) to analyze for manuscript?
    • 56 projects in CAMERA (many samples per project), quite a few are new this year
    • Story we'd like to tell: hit rare biome, find things you can't find without metagenomic data
    • Considerations: complex/diverse community, 454 sequencing (vs. Sanger), PCR and metagenomic data for comparison (check Josh's table or try CAMERA SQL query), published (?)
    • JL: Is a diverse data set needed to look at PCR bias? TS: competition for template higher/recurring in complex community
    • Do we want a low diversity community (acid mine, selected GOS sample)? Or simulate different levels of diversity w/ and w/out PCR bias?
    • ML: Total diversity or specific novel branches/clades? TS: Probably rarefaction type analysis, but will look for novel lineages
    • ML: Could higher sequencing error in 454 lead to false signal of greater diversity? TS: try to compare metagenomic and PCR with same sequencing method. Also, we would use quality filtering.
  • Tom and Steve will follow up on picking a data set or two
  • Tom and Sam will follow up on simulations

• Next time:
  • Josh
    • ranges
    • null models
  • James
  • others