SNP filtering by allele fraction

0

Entering edit mode

3.8 years ago

drowl1 ▴ 30

Hello,

I have a multi-sample VCF generated by GATK with an additional 'Allele fraction' (AF) annotation in the FORMAT field.

I wish to filter the SNPs in the VCF at the sample genotype-level using the AF in the FORMAT field and I want to remove sites with missing values (".") but without removing the whole variant.

I have tried the command GATK SelectVariant --select 'vc.hasAttribute("AF")' but that doesn't work.

Does anyone know of a simple way or a tool that could accomplish this?

Suggestions highly appreciated!

snp genome sequencing • 1.4k views

ADD COMMENT • link 3.8 years ago by drowl1 ▴ 30

1

Entering edit mode

Hi Could you please post an example of your table and the output you want to obtain after filtering.

ADD REPLY • link 3.8 years ago by hugo.avila ▴ 490

1

Entering edit mode

Hi,

I figured it out eventually. So the reason why I thought to remove sites with missing values was because when I tried a 0.5 filter threshold with the command - bcftools view -e 'FORMAT/AF[*] => 0.5', the whole variant was excluded for all samples even if only one of the sites failed the threshold or had missing values.

The command bcftools view -i 'MIN(FMT/AF)>=0.5' worked well by keeping the whole variant and just excluding sites that failed the threshold.

Thanks!

ADD REPLY • link 3.8 years ago by drowl1 ▴ 30

Login before adding your answer.