Test-drive the GATK tools and Best Practices pipelines on Terra
Check out this blog post to learn how you can get started with GATK and try out the pipelines in preconfigured workspaces (with a user-friendly interface!) without having to install anything.
HaplotypeCaller on whole genome or chromosome by chromosome: different results
I'm working on targeted resequencing data and I'm doing a multi-sample variant calling with the HaplotypeCaller. First, I tried to call the variants in all the targeted regions by doing the calling at one time on a cluster. I thus specified all the targeted regions with the -L option.
Then, as it was taking too long, I decided to cut my interval list, chromosome by chromosome and to do the calling on each chromosome. At the end, I merged the VCFs files that I had obtained for the callings on the different chromosomes.
Then, I compared this merged VCF file with the vcf file that I obtained by doing the calling on all the targeted regions at one time. I noticed 1% of variation between the two variants lists. And I can't explain this stochasticity. Any suggestion?