Filtering homozygous variants in vcf file with vcftools

[Seqql]

Hi everybody,

I'm trying to modify a vcf file to keep only homozygous variants.
I've tried using vcftools, but i can't get the syntax right.

I thought using the "keep-filtered" option of vcftools should be the right approach. Also I'm trying to find homozygous variants by the "GT 1/1" tag.

I've tried
vcftools --keep-filtered GT:1/1 --vcf in.vcf --out out.vcf
but it kept none.

edit:
vcf file looks like this:
c130956_g1_i1 50 . A C 5.49 . DP=2;RPB=-9.668049e-01;AF1=0.5049;AC1=1;DP4=1,0,1,0;MQ=33;FQ=-10.5;PV4=1,1,1,1 GT:PL: DP:GQ 0/1:34,0,17:2:22
c130956_g1_i1 62 . T A 49 . DP=3;VDB=5.867524e-02;AF1=1;AC1=2;DP4=0,0,2,1;MQ=30;FQ=-36 GT:PL: DP:GQ 1/1:81,9,0:3:16




What is the right syntax here?

[lindenb]

--keep-filtered is for the FILTER column

to answer this, I would use my tool https://github.com/lindenb/jvarkit/wiki/VCFFilterJS


java -jar dist/vcffilterjs.jar -e 'variant.getGenotype(0).isHomVar()' file.vcf

[Seqql]

Thank you, this is looknig very promising.

I got the filtered vcf file, but i've run into a problem.
I can't load the file into IGV because of the FILTER field, values like "11,30" aren't accepted. Looks like IGV only accepts one number with/out dots ("11.30" is ok).

Is there any quick fix for this I could apply to the code? If not i may just delete the whole column.

[lindenb]

11,30 is generated by my program (?) If yes, this is a locale problem:
try to set the LANG variable


$ export LANG=en_US.UTF-8
$ echo $LANG
en_US.UTF-8




or/and (?) specify it on the command line:

java -Duser.language=en -Duser.country=US (...)

[Seqql]

Good thought, it worked.
Thanks for the help

[werno]

Hello everybody,

can you help me how to filter heterozygous variants in vcf file?

thank in advance for answer

with best regards

werno