Skip to content

CHANGELOG.md

Latest commit

 

History

History
181 lines (126 loc) · 8.35 KB

CHANGELOG.md

File metadata and controls

181 lines (126 loc) · 8.35 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.9.4 - 2026-08-26

Fixed

  • Prevented Purple driver analysis failures by filtering somatic variants with missing reference or alternate allele depth (AD) values or missing read depth (DP).
  • Fixed ClairS-TO VCF sample naming so output files use the tumor sample identifier instead of the generic SAMPLE name.

See #27.

0.9.3 - 2025-10-27

Fixed

  • Updated PAVE processing of SAGE tumor-only calls to read only passing variants, preventing filtered variants from being included in downstream processing.

0.9.2 - 2025-07-22

Added

  • Added CPU and GPU execution paths for the DeepVariant call_variants step.
  • Added GPU-accelerated DeepVariant execution when the --use-gpus option is enabled.

Changed

  • Updated DeepVariant output handling to support the sharded TFRecord files produced by GPU execution.
  • Added separate resource allocations for CPU and GPU DeepVariant jobs.

0.9.1 - 2025-03-17

Added

  • Added HMF Tools SAGE somatic variant calling for tumor-normal and tumor-only samples.
  • Added ClairS-TO somatic variant calling for tumor-only samples.
  • Added DeepSomatic somatic variant calling for tumor-normal samples, with CPU and GPU execution support.
  • Added the --use-gpus option for GPU acceleration of supported workflow steps.
  • Added PAVE annotation and panel-of-normals filtering for SAGE tumor-only calls.
  • Added AMD and Intel processor detection for selecting a compatible BWA-MEM2 binary.
  • Added Zenodo citation and DOI information to the project documentation.

Changed

  • Replaced Strelka with DeepSomatic in the tumor-normal somatic variant workflow.
  • Updated the somatic caller ensemble:
    • Tumor-normal samples use Mutect2, Octopus, SAGE, MuSE, and DeepSomatic.
    • Tumor-only samples use Mutect2, Octopus, SAGE, and ClairS-TO.
  • Required tumor-only somatic variants to be supported by at least two callers.
  • Updated DeepSomatic to version 1.8.0.
  • Restricted DeepVariant and DeepSomatic WES processing to the configured capture regions.
  • Improved WES BED normalization to produce valid numeric scores and strand annotations.
  • Removed an additional Octopus genotype-level filter that reduced recall.
  • Removed the Mutect2 contamination-table filter after it was found to reduce recall.
  • Updated ClairS-TO resource allocations and region handling.
  • Adjusted Octopus resource allocations for observed memory requirements.

Fixed

  • Fixed DeepVariant and DeepSomatic WES calling with normalized BED input.
  • Fixed PAVE temporary-directory creation for SAGE tumor-only processing.
  • Fixed BWA-MEM2 execution on AMD Zen 4 processors by selecting the compatible AVX2 binary.
  • Disabled Snakemake HTML report generation after an upstream Vega dependency change caused report creation to fail.

Removed

  • Removed Strelka from the active somatic variant caller ensemble.

0.9.0 - 2024-06-28

Added

  • Added an optional GATK4 germline short-variant workflow enabled with --gatk-germline.
  • Added GATK HaplotypeCaller processing scattered across genomic regions.
  • Added cohort joint genotyping with GenomicsDBImport and GenotypeGVCFs.
  • Added SNP and indel Variant Quality Score Recalibration.
  • Added population-based genotype refinement and final cohort VCF generation.
  • Added support and documentation for the UGE execution backend.

Changed

  • DeepVariant remains the default germline caller; enabling --gatk-germline runs the GATK4 best-practices workflow in addition to DeepVariant.
  • Enforced use of Snakemake versions earlier than 8.0 because of incompatible upstream changes.
  • Expanded germline workflow resource configuration for SLURM and UGE.

0.8.0 - 2024-05-07

Added

  • Added UGE/SGE cluster execution through the --mode uge option.
  • Added scheduler-specific resource configurations for local, SLURM, and UGE execution.
  • Added support for a custom reference resource bundle through --resource-bundle.
  • Added automatic rewriting of bundled reference paths when a custom resource bundle is supplied.
  • Added generation of a matching FastQ Screen configuration for custom resource bundles.
  • Added reporting of detected Snakemake and Singularity versions.

Changed

  • Split DeepVariant into separate make_examples, call_variants, and postprocess_variants jobs for more efficient resource use.
  • Moved the original single-step DeepVariant rule to the deprecated workflow rules.
  • Updated memory calculations for UGE/SGE, where memory is allocated per CPU.
  • Reorganized cluster configurations by execution backend.
  • Improved Singularity cache handling outside scheduled jobs.

0.7.1 - 2024-02-06

Fixed

  • Moved FastQC intermediate processing to local scratch or the configured temporary directory to avoid locking and performance issues on GPFS filesystems.
  • Added local scratch allocations for raw-read and aligned-BAM FastQC jobs.

0.7.0 - 2024-02-02

Added

  • Added whole-exome sequencing support through --wes-mode.
  • Added support for custom exome capture regions through --wes-bed.
  • Added preparation, padding, sorting, compression, and indexing of exome capture BED files.
  • Added WES-aware DeepVariant, GLnexus, GATK, Qualimap, Manta, MuSE, Strelka, Octopus, and Mutect2 processing.
  • Added CNVkit copy-number analysis for paired tumor-normal WES samples.
  • Added Sequenza purity, ploidy, and copy-number analysis for paired tumor-normal WES samples.
  • Added WES pipeline documentation and CI dry-run coverage.

Changed

  • Restricted WES variant calling and downstream processing to configured capture regions where supported.
  • Selected WES-specific DeepVariant and GLnexus models when running in exome mode.
  • Disabled WGS-specific HMF Tools CNV processing for WES datasets.
  • Updated the project description from a WGS pipeline to a combined WGS and WES pipeline.

Fixed

  • Fixed parsing of SRA-style read identifiers when extracting flowcell and lane information.
  • Ensured normalized VCFs are compressed and indexed before interval-based GATK selection.

0.6.0 - 2023-12-06

First public release of genome-seek.

Added

  • Added comprehensive clinical whole-genome sequencing workflows for germline and somatic small variants, structural variants, copy-number variants, HLA typing, variant annotation, and quality control.
  • Added filtered Manta somatic structural-variant output using read support, allele fraction, coverage, and contig realignment.
  • Added Purple-derived somatic MAF files and cohort-level maftools summaries and plots.
  • Added population-frequency filtering of variants supplied to Purple.

Changed

  • Reworked somatic VCF normalization and merging across callers.
  • Replaced caller intersection with prioritized caller-aware variant merging.
  • Updated Purple to consume the merged somatic callset rather than separate caller-specific VCFs.
  • Updated Strelka output to include allelic-depth annotations required by downstream processing.
  • Standardized tumor-only Manta output naming with tumor-normal output naming.
  • Increased and tuned resource allocations for somatic calling, annotation, and CNV analysis.

Fixed

  • Sorted scattered Octopus VCF chunks before concatenation.
  • Improved filtering of Octopus calls to retain passing somatic variants.
  • Removed unnecessary splitting of normalized somatic VCFs into separate tumor and normal files.

0.1.0-beta - 2022-02-18

Added

  • Added the initial recommended scaffold for building a Snakemake pipeline.