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.
Mutect2 parallel problem
Dear GATK team.
I am using Mutect2 to call somatic mutation from tumor/normal paired sample. However after jobs running for 8 days, our server has been rebooted for some reason. Most of the jobs are done by more than 70%. For example, some jobs called variants at Chr14, some at Chr19, and it seems the variants calling are by chromosomes. May I ask is there a way to continue the unfinished part?
I used parallel option (-nct 4) and nonparallel option for the same jobs, but it turns out that the system used more time on communication among different threads, rather than actually speeding up. Parallel jobs are actually about 4 times slower than non-parallel jobs. Considering garbage collection problem, I added java -Xmx24G -XX:+UseConcMarkSweepGC -XX:ParallelGCThreads=4 ... in the command.
Could I submit Mutect2 for each chromosome, rather than whole genome? By submitting jobs for each chromosome can actually make the variants calling parallel.