VCF Header malformation

Question

[Mahdi&Christ]

Hi, I have a vcf file with 1 sample by 3 different variant caller.I want to do VQDR for it first and then merge those variants which are commen in at list between 2 variant callers.

_error:

There is a chrM that it's long is different from ref gene is used in vqsr proccess.

_question:

What should I do ?should I delete it from vcf?or just update the vcf header? Actually I updated header but there where another error : ID=GL number=.

I don't know what to do

And it also has chromosomes with chrUn_gl000

[GenoMax]

There is a chrM that it's long is different from ref gene is used in vqsr proccess.

You are not using the same reference or at least a reference where chrM (mitochondrial genome) is different in length.

[Jeremy Leipzig]

Why do you think VQSR is going to work with 1 sample?

Answer 1

[Aleksandra]

Hi! All of the above errors stem from a discrepancy between the reference genome used to create the VCF file and the sequence dictionary expected by the GATK VQSR. Manually editing the header does not solve the problem and will lead to invalid results. The correct way to proceed is to ensure that your VCF is fully consistent with the GATK dictionary. First, use bcftools norm to standardise the representation of variants in your VCF file. Then, specify the official .dict file from the GATK Resource Bundle when you apply gatk UpdateVCFSequenceDictionary. This will resolve any inconsistencies in the header