/CNValloc

A tool for estimating sequences of CNV alleles from multiple individuals

Primary LanguagePython

CNValloc

CNValloc

Requirements

  • Python 2.6 or 2.7
  • Modules: numpy, scipy, argtools

Description

A tool for estimating sequences of CNV alleles from multiple individuals. The allele ratio of each sample is also inferred.

A simple example

python cnvalloc estimate_alleles -v -K 4 examples/hist.txt

The result of estimation is emitted as JSON format.

Check performance for the number of haplotypes K = 1..10

$ parallel -k 'python cnvalloc estimate_alleles -K {} examples/hist.txt | python cnvalloc evaluate_alleles -r /dev/stdin -a examples/haps.txt' ::: {1..10}

  • examples/haps.txt : True haplotypes for evaluation

    • column 1: Sample id
    • column 2: Allele id
    • column n (n>2): The base of the allele at the n-2 th variant site

  • examples/hist.txt : Observed bases of sequence data at variant sites

    • column 1: Sample id
    • column 2: One of the 'ATCG' bases
    • column n(n>2) : The number of observed bases at n-2 th variant site

Workflow for BAM files

  1. Make pileup histograms from BAM files:

    $ python cnvalloc bam2hist {BAM file n} -r chr1:10000000-100010000 > pileups.n.txt

  2. Import the pileup files to a database such as sqlite3

  3. Select the variable sites to use with some criteria (e.g. minor_count >= 15 for any of the samples)

  4. Create an input file for the cnvalloc estimate_alleles by querying to the database

Future work

  • Consider variant types other than mutations
  • Write tools for step 2-4 of the above workflow

References

T. Mimori et al, 2015 BMC Bioinformatics "Estimating copy numbers of alleles from population-scale high-throughput sequencing data"

http://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-16-S1-S4