How to extract Homozygote variants froma VCF format?

Question

[Parimala Devi]

I am doing SNP analysis on whole genome saccharomyces cerevisiae. I want to segregate the homozygote variants from the heterozygote variants. How do I go about it?

Answer 1

[Gabriel R.]

Look at the GT field for the most likely genotype.

Answer 2

[Vivek]

For a diploid without multi-allelic loci

grep "^#\|1\/1" snps.vcf > hom-alt.vcf

[Parimala Devi]

I tried this, but I did find "0/1" still present in the vcf.

[Pierre Lindenbaum]

Show us your cmd-line. Is there only one sample? Can you find the string 1/1 elsewhere?

Answer 3

[Pierre Lindenbaum]

using my tool vcffilterjs

java -jar vcffilterjs.jar -e 'variant.getGenotype("SAMPLENAME").isHet()' your.vcf.gz > het.vcf

or use isHom() or isHomVar() or isHomRef()

or use GATK variant filtration

Answer 4

[Yahan]

snpSift is a useful tool if you want to look at a multisample vcf where all samples should be homozygote (or other types of queries):

for instance for four samples:

java -jar SnpSift.jar filter "countHom() = 4 & !(GEN[0].GT='./.') & !(GEN[1].GT='./.') & !(GEN[2].GT='./.')

& !(GEN[3].GT='./.')" -f my.vcf

It's unfortunate that it considers a no call (./.) also as a homozygote, but the !(GEN[0].GT='./.') ... part deals with that.