About ParabricksRelease Notes

4.5.1-1 Release Notes

View as Markdown

Highlights:

  • New features added in rna_fq2bam: single-cell mode, quant mode and batch mode.
  • --mutect-f1r2-tar-gz support for mutectcaller.
  • Added support for haplotype sampling in giraffe, enabling personalized reference pangenome creation per sample.
  • vcf.gz output support for all Parabricks variant callers.
  • New options for RNA-seq data in minimap2.
  • Updated base container to CUDA 12.9.

Improvements

Tool Updates

fq2bam and fq2bam_meth:

  • Optimized the performance of individual threads in the CPU thread pool during alignment.
  • Additional optimizations for GPU performance.
  • Improved error handling when using --gpuwrite option.

rna_fq2bam:

  • Added support for batch processing multiple input FASTQ files by repeatedly applying the --in-fq or --in-se-fq option.
  • Added support for batch processing multiple input FASTQ files by passing in the --in-fq-list or --in-se-fq-list option, where the input is a text file containing a list of paired-end or single-end FASTQ files.
  • Added support for single-cell sequencing mode, which can be enabled using the --soloType Droplet option along with additional necessary options including --soloCBwhitelist, --soloFeatures, --soloStrand, and --soloBarcodeReadLength.
  • Added support for quantification mode using the --quantMode option.

minimap2:

  • Added support for RNA options --minimizer-kmer-len and --forward-transcript-strand, which are recommended for use in mapping long RNA-seq data in splice modes.

mutectcaller:

  • Added argument --mutect-f1r2-tar-gz in mutectcaller to collect F1R2 counts for FFPE samples.
  • Added support for producing gzipped VCF files by specifying output file name with .vcf.gz extension.

haplotypecaller:

  • Added support for producing gzipped VCF files by specifying output file name with .vcf.gz extension.

deepvariant:

  • Added support for producing gzipped VCF files by specifying output file name with .vcf.gz extension.

deepsomatic:

  • Added support for producing gzipped VCF files by specifying output file name with .vcf.gz extension.

giraffe:

  • Added support for haplotype sampling, enabling personalized reference pangenome creation per sample.
  • Added real-time CPU and host memory usage monitoring via ‑‑monitor‑usage flag.

Bug Fixes

rna_fq2bam:

  • The flag --no-markdup is correctly processed.

fq2bam, fq2bam_meth, rna_fq2bam, and minimap2:

  • Fixed a bug where the --gpuwrite option crashed with “invalid device ordinal” or “invalid configuration argument” errors in rare cases.

collectmultiplemetrics:

  • Fixed a crash that occurred when the BAM file contained long reads.

  • Fixed a crash that occurred when the BAM header was not compatible with the reference.

For further information refer to the Parabricks datasheet.