New features, performance improvements and bug fixes by release.
NVIDIA Parabricks v4.3.1 has been released featuring new functionality for variant calling in somatic data and upgrades to the latest versions of industry-leading tools. Parabricks v4.3.1 expands NVIDIA’s emphasis on solving omics challenges with deep learning and continues accelerating genomics instruments.
What’s new
- Support for Google’s DeepSomatic in short-read sequencing
- Upgraded version of DeepVariant (v1.6.1)
- Upgraded version of Minimap2 (v2.26)
- Bug fixes and performance improvements for haplotypecaller, mutectcaller, rna_fq2bam, fq2bam, and fq2bamfast
Check out more in documentation
Read technical blog
Would these bug fixes address the low recall that was observed here?
I would love to use parabricks as the run times are fantastic but it is missing/not-calling a lot of variants.
Also when you release a new version of parabricks, would it be possible to release benchmarking information for the germline-related workflows as well? It would be awesome if you could run and benchmark germline workflows against truth-set (say GIAB sample HG002), and provide the results as part of the release notes or in your documentation for the relevant tool/workflow.
With that being said, thank you so much for continuing to improve these tools! I am super excited for the future of parabricks. I know porting these tools is a massive undertaking, and your team is doing an extremely exceptional job!