Re: [Vcftools-help] Can I subset my vcf file (with mulitple individuals genotyped) by SNPs/indels/other variants

[Petr D. <pd...@sa...>]

Hi,

there is vcf-subset script which can subset VCF files by SNPs or indels.
It would be straightforward to do the same for other variants as well.

On Thu, 2011-02-17 at 11:34 +0100, Mao Jianfeng wrote:
> Dear vcftools listers,
> 
> I just do NGS analysis depending on VCFTools. And, I would like to
> conserve the consistency of  various variant definition and statistic.
> 
> Sorry for my simple question.
> 
> I have got my vcf file annotated. So, now I want to count the number
> of different variants (say, SNPs/indels/other variants in my case)
> overlaped cds/intron/exon/genes. But I do not know how to identify the
> distinct variants when counting. I guess I may need to divide my vcf
> file (with mulitple individuals genotyped) into three different
> subsets each of which corresponding to SNPs, indels or other variants.

you could use vcf-annoate to add information about cds/intron/exon/genes
into INFO column of your VCF file. Then vcf-stats script can be then
used to get the numbers for each category.

> I think this should be easy to implement by using VCFTools, because
> vcf-stats can identify each of the different variant already. So,
> could drop me lines of code/scripts to do this subset by variants?

Not tested, but something like this should give you a good start:

$vcf = Vcf->new(file=>$file);
while (my $line=$vcf->next_data_array)
{
   my $ref = $$line[4];
   my (@alts) = split(/,/$$line[5]);
   for my $alt (@alts)
   {
      my ($type,$len,$ht) = $vcf->event_type($ref,$alt);
      # More info about return type in `perldoc Vcf.pm`
   }
   ..
   if ($ok_to_print)
   {
      print $vcf->format_line($line);
   }
}

Petr


> I am a population genetics and I have few bioinformatic programming
> skills in my own side.
> 
> I expect your reply, thank you in advance.
> 



-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.