VCF with SNP- count number of 100% missing genotypes

Question

[misterie]

Hey,

I am using a VCF Tools for SNPs filtering. I need to count (or extract) a variable which have 100% missing genotypes (I have multisample VCF).

In VCF I know a command --max.missing, but I can define only a max missing, without a min number.

It is important for me, because I have 1 big VCF with 8 breeds' SNP's, and I would like to get information, how many SNP's are exactly in each breeds.

Thank you in advance!

Answer 1

[Jeremy Leipzig]

genomic loci which have 100% missing coverage (or all reference calls) would not show up in a VCF anyway

Answer 2

[chrchang523]

You can dump the list of SNPs which are less than 100% missing, and then --exclude that list.

[misterie]

Which command allow me to exclude this 100% missing data?

Answer 3

[misterie]

Look up. I told I am extracting from multisample file each breed and it generate empty variable. In my VCF file I have many breeds. I am extracting just one, and this file contain a null variable, but other breed in the same position, SNP was detected.

Answer 4

[Pierre Lindenbaum]

using vcffilterjdk: http://lindenb.github.io/jvarkit/VcfFilterJdk.html

 java -jar dist/vcffilterjdk.jar -e 'return !variant.getGenotypes().stream().anyMatch(G->G.isCalled());' in.vcf

[misterie]

I can not use it on my server. I can not install new softs. Exist any other option, to get this 100% missing data, using any bash, vcftools, beagle or another?

[Pierre Lindenbaum]

I can not install new softs

you don't have to install it on the server, you can always install it just for you. (and in fact, it is installed where it's compiled... )

[misterie]

but it will be impossible to run this on my files, which are on the server. I can't download it, because its too big file :(

[Pierre Lindenbaum]

wget -O - "http://a.c.d/path/to/file.vcf.gz" | gunzip -c |  java -jar dist/vcffilterjdk.jar -e 'return !variant.getGenotypes().stream().anyMatch(G->G.isCalled());'