Extract only first individual from vcf files

Question

[User000]

Dear all,

I have multiple vcf files like this, each containing 3 individuals. I would like to extract only the first individual from each vcf and save each as separate vcf file. Is this possible in batch without specifing the individual name as I have many vcfs?

#CHROM  POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  NAME1   NAME2   NAME3
chr1    1666251 .   G   A   139 MQ  BRF=0.46;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-143.73,0.0,-79.13:20:99:100:60 0/0:0.0,-12.67,-236.1:12:99:117:2   0/0:0.0,-16.47,-244.0:13:99:129:1
chr1    2408213 .   T   G   164 QD  BRF=0.05;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-21.02,0.0,-55.12:44:99:77:24   0/0:0.0,-13.94,-147.2:5:99:58:0 0/0:0.0,-8.73,-90.8:3:87:37:0
chr1    2408232 .   T   G   122 QD      BRF=0.04;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-16.82,0.0,-33.02:51:99:57:23   0/0:0.0,-10.38,-111.1:9:99:37:0 0/0:0.0,-5.92,-60.6:8:59:25:0

Expected output:

   #CHROM   POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  NAME1
chr1    1666251 .   G   A   139 MQ  BRF=0.46;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-143.73,0.0,-79.13:20:99:100:6
chr1    2408213 .   T   G   164 QD  BRF=0.05;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-21.02,0.0,-55.12:44:99:77:24
chr1    2408232 .   T   G   122 QD      BRF=0.04;FR=0.1667  GT:GL:GOF:GQ:NR:NV  1/0:-16.82,0.0,-33.02:51:99:57:23

[cpad0112]

simple:

$ cut -f1-10 -d$'\t' input.vcf

For many:

$  parallel --plus --dry-run  cut -f 1-10 -d\$\'\t\' {} ">" {.}_out.vcf ::: *.vcf

[Pierre Lindenbaum]

sure, but it could be vcf.gz or bcf and bcftools will fix AC.

[cpad0112]

user can use zcat for gz.

[zx8754]

Is the first individual's ID always "NAME1"? Something like this:

bcftools view -s NAME1 filename.vcf.gz

[User000]

no, that's the problem..

[zx8754]

I see, then go with Pierre's solution.

Answer 1

[Pierre Lindenbaum]

find dir1 dir2 -type f -name "*.vcf.gz" | while read F
do
   bcftools query -l "${F}"| head -n 1 > samples.txt
   bcftools view --samples-file samples.txt -O z -o "${F}.firstsample.vcf.gz" "${F}"
done

[User000]

thanks, could you please explain a bit the code? what is dir1 dir2 i.e?

[zx8754]

dir1 and dir2 are your directories where you have VCFs. Find vcfs, loop through them using while, get the first sample ID, then subset.