Feature request: --static-quantized-quals for applybqsr

Healthcare & Life Sciences Parabricks
1

Hi Parabricks team,

We’d like to request support for GATK’s --static-quantized-quals option in pbrun applybqsr.

Context. We use Parabricks through nf-core/sarek, which runs fq2bam and then applybqsr (added in sarek’s dev branch) to write a recalibrated BAM/CRAM per sample. That file is kept long-term and used by several callers, including non-GATK ones such as Strelka and Manta that can’t apply a BQSR table on the fly.

Problem. Without quality quantization, recalibrated files are much larger than their input. With GATK ApplyBQSR run without quantization, our recalibrated WGS CRAMs were roughly 1.3–3.5× the size of the duplicate-marked CRAMs, and we’d expect the same from pbrun applybqsr, whose documented GATK equivalent is ApplyBQSR without quantization. The Broad reference pre-processing workflow (gatk4-data-processing) instead runs ApplyBQSR with --static-quantized-quals 10 --static-quantized-quals 20 --static-quantized-quals 30, which in our experience keeps recalibrated files close to input size. pbrun applybqsr has no equivalent option (checked against the 4.7.1 tool reference), so there’s currently no way to produce output matching that workflow.

Request.

  • Add --static-quantized-quals (repeatable, same semantics as GATK ApplyBQSR) to pbrun applybqsr.

Since haplotypecaller already implements this quantization, we hope it’s a small addition for applybqsr. Happy to test a pre-release build on our data.

Thanks!

2

Hello,

Thank you so much for reaching out!

We will evaluate the importance of this feature and do prioritization accordingly for the future release. If you could tell us who you are and share more details on your use case, that would be helpful for us to prioritize. Feel free to message me.

Thanks!