Purpose

splitaake is a reasonably easy method for demultiplexing Illumina reads using either Hamming distance or Levenshtein (edit) distance sequence tags. A the moment, Hamming distance is the default, which is similar to the approach Illumina uses in their own software. splitaake differs from the Illumina software in that it can demultiplex sequence tags of many lengths, rather than just the "standard" TruSeq index length of 6 nucleotides.

Notes

splitaake is under development. This means that it may break, be a pain to get running, etc. I'll improve as I have time. Please feel free to suggest contributions/changes/additions.

Design

splitaake is currently designed as a single-core application meaning that it does not parallelize the process of demultiplexing. After a number of tests, I've found that for most files, a single core approach is reasonably fast (and sometimes faster) than multi-core options, particularly when you wish to work with gzipped fastq files.

I'm still testing additional ways of demultiplexing data in parallel. Hopefully more on this front soon...

Dependencies

• seqtools ("working" branch):

  pip install git+git://github.com/faircloth-lab/seqtools.git@working
  

• jellyfish (at the moment - quite fast Hamming implementation):

  pip install git+git://github.com/sunlightlabs/jellyfish
  

Running

• generate a config file mapping indexes to filenames. This file is named map.conf, as used below:

  TruSeq1:ATCACGATCT
  TruSeq2:CGATGTATCT
  TruSeq3:TTAGGCATCT
  TruSeq4:TGACCAATCT
  TruSeq5:ACAGTGATCT
  TruSeq6:GCCAATATCT
  TruSeq7:CAGATCATCT
  TruSeq8:ACTTGAATCT
  TruSeq9:GATCAGATCT
  TruSeq10:TAGCTTATCT
  TruSeq11:GGCTACATCT
  TruSeq12:CTTGTAATCT
  

• run splitaake:

  python splitaake.py L007_R1.fastq.gz L007_R2.fastq.gz L007_R3.fastq.gz map.conf --section taxa
  

• this will identify your reads and create a directory dmux containing your reads in interleaved, fastq, gzip files, like so:

  dmux/
      TruSeq1.fastq.gz
      TruSeq2.fastq.gz
      TruSeq3.fastq.gz
      ...