Version 1.2.0

Changes in this release:

• Singularity container support for all rules. In fact, singularity is now the preferred method of executing the pipeline. This allows much better isolation of run environments.
  • Versions of tools are the same as in the (now legacy) conda environments, bar one: fastqc. When using singularity, fastqc uses version 0.11.7, as opposed to version 0.11.5 when using conda. The differences are minor, with the largest difference being that the per-tile statistics for Novaseq runs only work on 0.11.7.
• Gather steps now use bcftools in stead of GATK's CatVariants. This is considerably faster, and less memory-intensive.

Version 1.1.1

Changed in this version:

• Set memory limits on markduplicates step, which fixes a memory error for WGS runs of >1.5B reads per sample
• Remove build numbers from conda environment files, as those get retired.
• Fixes conda environment files containing matplotlib
• Per-exon coverage metrics will have slightly lower mean_gq values, as vtools now calculates mean gq over the raw P-values.

Version 1.1.0

Changed in this version:

• Added testing through pytest-workflow and gitlab-ci
• Remove superfluous --OUTPUT_DIR configuration value.
• Remove superfluous --JAVA_EXE configuration value
• Fixed an issue where seqtk rule created dangling softlinks.
• Bump required snakemake version to at least 5.2.0