Output a vcf restricted to variant sites listed in a file

Question

[modernsynthesis]

I'm looking for a tool that will output a vcf restricted to variant sites (not genomic intervals) listed in a tab-delimited file. I understand GATK SelectVariants can do this, but I'd prefer an alternative tool that doesn't reorder samples.

Answer 1

[Chris Miller]

Not quite sure I understand why this won't work, as long as your bed file is composed of 1bp variant sites and not intervals.

bedtools intersect -a my.vcf -b my.bed -wa >output

If you have something different in mind, then example data would be helpful.

[modernsynthesis]

this works nice, but is there a way to keep the vcf metadata so i can pipe to bcftools?

[Chris Miller]

Do you mean the header? A couple of redirects will get you there:

cat <(head -n 10000 my.vcf | grep "^#") <(bedtools intersect -a my.vcf -b my.bed -wa) | downstreamToolX

[modernsynthesis]

great, thanks!

Answer 2

[Pierre Lindenbaum]

Convert your vcf / target region to bed. (For a VCF it's something like:

grep -v "^#" your.vcf | awk -F '\t' '{printf("%s\t%d\t%d\n",$1,int($2)-1,int($2)+length($4)-1);}' > input.bed

and then use bcftools view with option --regions-file

or

bedtools intersect..