This is a follow up to the discussions on the PR; and how to best do this.
There are some recommendations on whether to parallelize at the engine or tool level from the GATK, and also some other tricks that people may wish to use to speed up processing (e.g. : dedup/ BQSR/call variants per chromosome (or some other criteria), or skip the PrintReads stage in BQSR ...etc)
I'm adding this as an issue, with some discussions on the pull request above
This is a follow up to the discussions on the PR; and how to best do this.
There are some recommendations on whether to parallelize at the engine or tool level from the GATK, and also some other tricks that people may wish to use to speed up processing (e.g. : dedup/ BQSR/call variants per chromosome (or some other criteria), or skip the PrintReads stage in BQSR ...etc)
I'm adding this as an issue, with some discussions on the pull request above