Filtering Variants using Snpsift filter and SnpEff

Question: Filtering Variants using Snpsift filter and SnpEff

22 months ago by

Hi,

I have a bunch of variants in .VCF file which I got through the unified genotyper tool. Now I want to filter only the "Homozygous" (100% alternate allele frequency) variants. But the Snpsift filter and SnpEff tools do not have any option to change what on basis the variants are annotated as "Homozygous" and Heterozygous". At least is there are way to filter out the variants based on frequency of alternate non-reference alleles.

In my case I need only the variants with 100% of alternate non-reference alleles. Can someone help me on this?. It would be a great help.

Thanks in Advance.

snp galaxy • 1.9k views

ADD COMMENT • link •

modified 22 months ago by Guy Reeves • 1.0k • written 22 months ago by braveen.joseph • 20

22 months ago by

Guy Reeves • 1.0k

Germany

Guy Reeves • 1.0k wrote:

I find it hard to follow what you are trying to do. Do you want to output a file with only those sites were all samples are 1/1? If this is what you want to do this try ((countRef() < 1) AND (countHet() <1)) To be honest I have not tested it (and I can see that any sites with missing data would make it into the output file). Commands explained on http://snpeff.sourceforge.net/SnpSift.html

ADD COMMENT • link modified 22 months ago • written 22 months ago by Guy Reeves • 1.0k

Hi,

Thanks a lot for the reply. I think what you said above is working. What i actually wanted is... let's say there are 30 reads for a nucleotide site in the WGS alignment. If a mutation is present on a site, I wanted only the sites where the mutation is present in all 30 reads in my output file (100% alternate allele frequency). i tested your code and cross checked few variants with the alignment. It looks good.

Still i'm finding it really hard to understand the documentation of the tool (The code). Will you be able to explain brifley in layman terms. Thanks again for the great Help.

ADD REPLY • link modified 22 months ago • written 22 months ago by braveen.joseph • 20

22 months ago by

Guy Reeves • 1.0k

Germany

Guy Reeves • 1.0k wrote:

(countRef() < 1)= keep site if the the total number of samples with 0/0 is less that 1 ie zero. (countHet() <1)= keep site if the the total number of samples with 0/1 is less that 1 ie zero. (AND ) is used ensure that only sites where there are no 0/0 AND no 0/1 ie sites with only 1/1
If by '30 reads' you mean 30 samples. You could also filter out sites with any missing data by

((countRef() < 1) AND (countHet() <1) AND (countHom() > 29))

Cheers Guy

ADD COMMENT • link written 22 months ago by Guy Reeves • 1.0k

Please log in to add an answer.

Similar posts • Search »