VCF-Annovar question

[quantrix]

Hi there,
I am relatively new and working through stuff. I have looked around and high and low for an answer and could not find one. So here goes.

I am trying to use annovar to convert a VCF file for analysis. When I run convert2annovar.pl, I get the following error.

Error: invalid record in VCF file: the GT specifier is not present in the FORMAT string: <chr3 41265950 . ATTCTTTT ATTTT 68.5 . INDEL;IS=10,0.072993;DP=136;VDB=7.627334e-21;AF1=0.5;AC1=1;DP4=8,5,2,20;MQ=44;FQ=68.9;PV4=0.0017,1,1,1 PL 106,0,107>

Needless to say, I examined my .vcf file and here is a field
chr3 41274764 . C A 222 . DP=4288;VDB=1.423052e-28;AF1=1;AC1=2;DP4=0,0,2246,2027;MQ=44;FQ=-282 PL 255,255,0

It is missing the "GT" tag.
My question is, at what step do I introduce the GT tag.

Is it with Bowtie alignment or Samtools mpileup or do I have to use separately vcftools to get the GT format in there?

I know this is a basic question, but I am trying to figure stuff out here. Thanks for the favor of a reply. I felt the vcf documentation was not quite clear about this.

[vivek_]

It should be generated in the VCF file produced after samtools mpileup variant calling stage. WHich version of samtools are you using?

I see this from the samtools webpage:

The VCF file produced by BCFtools does not strictly conform the VCF spec. For example, the GT genotype information is not always present because for the purpose of BCF, GT is unnecessary and takes disk space. In addition, GT is not the first as is required by the VCF spec. This can be fixed by the bcf-fix.pl script that comes with the source code package, and will be fixed in future (fixed in r880+).

http://samtools.sourceforge.net/mpileup.shtml

[quantrix]

Hi Vivek,
Thanks for the quick reply. I am using samtools-0.1.19. I shall look more into this and try to figure out. I would be glad for any further guidance as well.
Regards

[vivek_]

I'm not aware of your variant calling methods but I think the basic idea should be using samtools mpileup on the bam file -> output BCF -> use BCFtools to generate a VCF file (Here use -g option to generate GTs)

You could also use GATK's unified genotyper on the BAM, which produces a VCF output by default.