CombineGVCFs takes forever if there are no calls in one g.vcf
we have one sample which produced only ~2000 mapped reads and therefore got no calls at all from HaplotypeCaller. Since we are doing the whole pipeline at once, we merged all per sample g.vcf of that run into one to do GenotypeGVCFs. In most runs this taks a few hours. In this case it took 2.5 weeks. As I removed the g.vcf of this sample it was done in 6 hours.
Why does CombineGVCFs takes so much longer if there is one file without calls? I attached a g.vcf of chromosme 2 as a txt file (since g.vcf was not possibe).
Looking forward to your ideas.