Home

vafCorrect: unbiased variant allele fraction correction

If an individual has more than one tumour sample it is helpful to perform comparative analysis of variant data on multiple tumour samples in unbiased manner.

vafCorrect facilitate unbiased variant allele fraction calculation for SNP and indels for union of variant sites in the set of user defined group of samples.

vafcorrect filter reads with flags NOT_PRIMARY_ALIGN, VENDER_FAIL, UNMAPPED, SUPP_ALIGNMENT however uses any unmapped reads whose pair is mapped within the vicinity [ defined by library insert size] of indel site, resulting reads were aligned using exonerate to original reference sequence and alternate reference sequence [ created by inserting indel variant at given reference location].

Exonerate output is parsed to count fraction of reads aligned to original reference [WTR: Wild type reads] and alternate reference sequence[MTR: Mutant reads].

Reads that are mapped with equal identity score on reference and alternate sequence were reported as ambiguous [AMB] reads while reads that are present at the variant location however doesn’t map to either reference sequences wee categorised as unknown reads[UNK].

For substitution we discard reads that are flagged as NOT_PRIMARY_ALIGN,VENDER_FAIL,DUP_READ,SUPP_ALIGNMENT.

Bio::DB::HTS is used to get unbiased pileup results.

VAF: variant alleler fraction is calculated as: VAF = MTR/(WTR+MTR)

1. Handling ambiguiously mapped reads

2. Example command line

cgpVaf.pl -d testdir -o testdir/output  -g testdir/genome.fa -a indel -e .vcf.gz \
  -nn <normalSample> -tn <tumourSample>

User can restrict analysis to set of locations using custom bed file

Example input bed file [optional]:

#chr    pos ref alt
1       16894612        T       G
1       16900912        G       A
1       16901539        G       T