[FEATURE] GATK GenotypeGVCFs option?

[mvelinder]

It doesn't appear GenotypeGVCFs included in --tools? Could it be included in a future version? It's an important step for some larger scale studies. GATK documentation and description: https://gatk.broadinstitute.org/hc/en-us/articles/360037057852-GenotypeGVCFs

Thanks!

[mvelinder]

Nevermind, it seems this is included here https://nf-co.re/sarek/2.7.1/output#gatk-genotypegvcfs Sorry for the confusion

[mvelinder]

So while it is listed in the link I gave previously, https://nf-co.re/sarek/2.7.1/output#gatk-genotypegvcfs - it does not work with --tools . What is the usage for GenotypeGVCFs then?

[maxulysse]

Hi @mvelinder --tools haplotypecaller --generate_gvcf should do the trick.

[mvelinder]

Thanks @maxulysse - from the GATK tool, the output is "A final VCF in which all samples have been jointly genotyped." https://gatk.broadinstitute.org/hc/en-us/articles/360037057852-GenotypeGVCFs

My latest run using --generate_gvcf (starting from 3 sets of FASTQ files as a trio exome) did indeed result in a g.vcf.gz for each sample. Have these individual GVCFs been jointly genotyped but just not merged into a single multi-sample output VCF? Would it be appropriate to just bcftools merge the 3 individual GVCFs to create a multiple sample joint-genotyped VCF?

Appreciate your help and clarification!

[maxulysse]

Hi @mvelinder.
At the moment in sarek the g.vcf are not merged yet, so each g.vcf file you have is from one set of FASTQs.
I'd look more into GenomicsDBImport for merging the resulting g.vcf files.
Hopefully it's something that we'll be able to add to the next sarek release.

[FriederikeHanssen]

the GATK recommended workflow for joint genotyping is added here #595 including genomicsdbimport, genotypegvcf, and vqsr