Moore Notes 2 18 09
From OpenWetWare
Jump to navigationJump to search
Group Call
- GBMF needs more detail on UCSF expenditures for annual budget
- Eric Matson (Berkeley) and Robin Kodner (from Ginger Armbrust's Lab, UW) are visiting Eisen lab
- they are working on metagenomic phylogenetics
- Update from James
- spatially explicit taxanomic predictions (building on PNAS paper)
- stochastic processes and allometric scaling
- unified neutral theory (taxa area relationships/distance decay, two-point correlations)
- JE: How will the predictions be tested with metagenomic data?
- problems: measuring area, can't count everything (and undersampling is a function of area/volume)
- Finley Science paper: protist analysis that uses every species and whole environment as a data point
- Josh is working on similar things, and we will follow up in a separate call
- Taxa-individual relationships (rarefaction)
- Distance-decay may be more feasible (undersampling is still an issue, may be different for z)
- Simulations
- Which taxa? How many?
- How to choose taxa? Are they in AMPHORA?
- reliably placed ones
- Dongying's phylogenetic distance (evenly sampled from tree) - simple case for tree building
- to look like metagenomics data (clusters)
- include archaea?
- Complete sequences vs. complete with fragments
- What gene(s)? Need full length alignments - complete data answer
- 16S rRNA (metagenomic vs. full length)
- RecA or some faster evolving gene (can build HMM with ZORRO)
- 2 AMPHORA markers (31 HMMs for key markers), e.g. longest and shortest
- DNA vs. codon vs. amino acid
- Answers depend on the use of the simulated data
- How to make alignments?
- What does pool of reads look like?
- to start, one read per taxon not overlapping - does it get placed on the full tree correctly?
- same, but with overlapping reads
- mulitple reads non-overlapping per taxon - do they place the reads together?
- mix of rare (not overlapping) and abundant taxa (overlapping)
- Problem: how to assess trees when leaves vary between them?