Cbcb:Pop-Lab:Ben-Report

From Cbcb
Revision as of 15:38, 16 February 2009 by Langmead (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Recent work.

Snapshot:

I'm working hard on getting paired-end support into Bowtie. I've settled on the best technique for doing this and I've implemented it in the -v 0 and -v 1 (no mismatches, one-mismatch resp.) alignment modes. I am now testing it, hoping to release a new version of Bowtie in the next couple days.

Outline of future work:

  • Bowtie
    • Paired-end support
      • Settled on a "mixed-mode" paradigm: find anchor mates in BW space, find opposite mates in reference space
        • Index becomes larger to accommodate the reference text (2.2 GB -> 2.9 GB for human)
        • Speed is about 2x slower than if all the mates were aligned in single-end mode
        • Default mode is "best-effort" - if you want Bowtie to bend over backwards to report all non-concordant pairs, that will be much slower but it can be activated from the command line
    • Gapped alignment
    • Mapping qualities
      • A la Maq
    • Color space
  • Crossbow
    • Already have some results using some 1000-Genomes data: about 1 hr to align about 14.5x coverage for human
    • Would like to do another run that includes paired-end reads and for which there is a trusted set of called SNPs