> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/clara/parabricks/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/clara/parabricks/_mcp/server.

# 4.3.0-1 Release Notes

**Highlights:**

* New tool, [fq2bam\_meth](/tool-reference/tools/fq-2-bam-meth) for accelerated DNA methylation analysis.
* Germline resource mode and force calling mode are supported in [mutectcaller](/tool-reference/tools/mutectcaller).
* Support for writing CRAM files using queryname-based sorting has been added into bamsort.
* Parabricks toolkit (v4.2) is now fully supported on Grace Hopper, refer to [NVIDIA Grace CPU Platforms](/get-started/nvidia-grace-cpu-platforms).
* Performance improvements in [germline](/tool-reference/tools/germline-gatk-germline-pipeline) and [deepvariant\_germline](/tool-reference/tools/deepvariant-germline) running on DGX H100.
* [deepvariant](/tool-reference/tools/deepvariant) version 1.6 updated.
* [minimap2](/tool-reference/tools/minimap-2) performance improvements and bug fixes.
* fq2bamfast performance improvements and bug fixes.

## New Tools

New tool for bisulfite sequencing data [fq2bam\_meth](/tool-reference/tools/fq-2-bam-meth)
based on [bwa-meth](https://github.com/brentp/bwa-meth/). Our tool,
[fq2bam\_meth](/tool-reference/tools/fq-2-bam-meth), implements compatible pre- and
post-processing around BWA MEM for DNA methylation analysis. It uses the same
accelerated alignment code as is used in fq2bamfast to
produce fast and accurate alignment.

## Improvements

### Tool Updates

[bamsort](/tool-reference/tools/bamsort):

* Supports CRAM file write on queryname-based sorting. It auto-detects cram file extension on output file.

[mutectcaller](/tool-reference/tools/mutectcaller):

* Adds the following new options:

  * `--mutect-germline-resource`
  * `--mutect-alleles`
  * `--force-call-filtered-alleles`

[deepvariant](/tool-reference/tools/deepvariant):

* Updates to match the baseline version v1.6.

[minimap2](/tool-reference/tools/minimap-2):

* Reduced reader buffer size to shorten the time it takes to start processing.

fq2bamfast

* Speed improvements.
* Added support for BWA MEM options: `-B` (values up to 15), `-T`, `-L`, and `-U`.
* Added support for reads longer than 500 bp using CPU recovery mode (note that speed will be slower and memory usage will be higher). Set `--max-read-length` to the desired max read length for the FASTQ filter.

### Improvements Spanning Multiple Tools

* Better messaging in filehandle when reading index files to avoid user confusion.
* Better error checking when reading FASTQ files: checks that *each* FASTQ read name line starts with '@'.

## Bug Fixes

* [mutectcaller](/tool-reference/tools/mutectcaller) and [haplotypecaller](/tool-reference/tools/haplotypecaller): Fixed a random segfault in bamOut mode.
* [haplotypecaller](/tool-reference/tools/haplotypecaller): Fixed an "allele out of index" bug in gvcf mode.
* [haplotypecaller](/tool-reference/tools/haplotypecaller): Fixed an GPU shared memory overflow bug in gvcf mode.
* [haplotypecaller](/tool-reference/tools/haplotypecaller), [mutectcaller](/tool-reference/tools/mutectcaller) and [deepvariant](/tool-reference/tools/deepvariant): Fixed a wrong return value 0 when the run fails.
* [minimap2](/tool-reference/tools/minimap-2): Corrected banner name when running minimap2.
* [minimap2](/tool-reference/tools/minimap-2): Fixed overflow issue in postsort.
* [minimap2](/tool-reference/tools/minimap-2): Removed max read size requirement that made some inputs unable to run.
* fq2bamfast: Fix for edge case (related to multiple hits) with `-C` option in BWA MEM to copy auxiliary tags from FASTQ comments.

For further information refer to the
[Parabricks datasheet](https://resources.nvidia.com/en-us-hc-genomics/healthcare-genomics).