Vcf Format And Filtering

Question

[Ashutosh Pandey]

Hi All,

I have generated variant data using samtools mpileup and bcftools. Now I am using vcftools to filter out variants based on my criteria. The genome I am working on is an inbred mouse so I should only see homozygous variants. Also, vcftools provide you with bunch of other parameters to filter your variant data. My vcf file consists of only one sample.

Questions:

1) Should I use AF1=1;AC1=2; tag in vcftools to select the homozygous mutations OR is there other better way to do it? OR Should I be using little easy criteria such as AF1>=.95 . I also have a DP4 tag info that gives you the read counts for ref and non-ref. But I assume AF1 is calculated based on DP4 information only.

2) I have around 80X data. I am using vcftools to filter my variants. I have kept most of the parameters to their default settings but have no idea about two parameters. a) W, GapWin INT Window size for filtering adjacent gaps [3] b) a, MinAB INT Minimum number of alternate bases (INFO/DP4) [2]

Can someone please explain me these two parameters. I have no idea what they are talking about.

Thanks

Answer 1

[Ashutosh Pandey]

Hello Everyone,

1) Just found out that the FQ tag in the vcf format will have positive values for heterozygous variants and vice-versa for a single sample. Also, the PL tag can also be used as it has the Phred based genotype scores as RR,RN,NN. For a Homozygous mutation the PL tag should have values like 255,255,0.

[potassiumiodide0990]

^^ you mean bcf format right? Coz I didnt find FQ tag in vcf format...

Answer 2

[Sean Davis]

Take a look as snpSift or vcftools. You can try filtering on QUAL (>40, for example) and a homozygous variant calls.

Answer 3

[Erik Garrison]

I described a method to do these kinds of operations using vcflib here: Filtering Vcf File