1.3.2

[1.3.2] - 20250829

Housekeeping

• Cleaned up variable syntax for clarity and consistency
• Ensure no memory issues on laptops for pixi run test (hash based deduplication memory can spike on FASTP)
• Aligned index kmer sizes in params.
• Converted slurm example script to a pipeline agnostic template

Vg command updates

• For complex graphs, vg index may cause memory issues. Users can switch on --noNestedDistance to add this option to the initial vg indexing commands and produce a limited version of the distance index. Note this currently wouldn't help for the pangenome-map module, in which case we'd consider upgrading that module to a process_high label

• vg filter - removed unused ext.args

1.3.1

[1.3.1] - 20250819

Housekeeping

• Updated dag that was supposed to be in last release
• Merged PR #4 (fastqc module now allows for fq.gz or fastq.gz extensions)

Software updates

• pixi.toml 1.1 Apptainer: 1.3.6 -> 1.4.1

• pixi.toml 1.1 nextflow: 24.10.4 -> 25.04.6

  • Nextflow update required syntax change for publishing outputs in main.nf + removal of topic feature flag in grave.nf. Small tweaks made to initialise subworkflow & README.md to reflect update

• pixi.toml 1.1 vg: 1.63.0 -> 1.67.0 (updated bioconda recipe)

• Module containers: vg version to 1.67.0 in containers

Configs

• Moved params to nextflow.config following nf-core best practices
• Same for base profiles. This means there is now an apptainer profile, replacing containers.config. Other base profiles added are debug and singularity. README.md updated to reflect this

Documentation

• Extra guidance/background on pixi usage, general improvements

Pipeline debugging/tracing

• Added --tracing parameter to enable comprehensive reports for the workflow, including timeline, report, trace, and dag files (stored in ${projectDir}/tracing). Off by default.

Bug fixes

FASTQ_MERGE_DEDUP: fixed potential issue that would arise if sample names contain ".", and generally cleaned the library count variables

FASTQ_MERGE_DEDUP: fixed bug for modern samples with multiple libraries (only first library was taken due to syntax error) #3

1.3.0

• README.md updated with workflow output details and more information on genotyping and variant calling tools

• Most relevant, now explicitly point out the difference between the genotyping and variant calling tools

• Implemented storeDir for outputs that need to be run once (snarls, hapl index, filter index). This is helpful when multiple users share a file system for storing graphs. Processes won't rerun even on fresh repo clones

• These now store to a new folder next to the graph based on the graph basename, e.g.: human.gbz, indexes are stored in: human_indexes

• To ensure other stored indexes are not picked up by other stores, each index category goes into a specific subdirectory

• Linked to adding storeDir, temporarily disabled topic channel version reporting from 3 modules that use it (nextflow-io/nextflow#5785)

• Linear references extracted from the graph are now included as output files in results/linear_references

• ext.args template added to each module, and placeholder module config file added conf/modules.config

• changed names of genotyping modules, channels, params etc. to increase their descriptive clarity - e.g. vg-graph-call -> vg_deconstruct, vg-map-call -> vg_genotype

• for clarity genotyping outputs go into a genotyping folder, not variant_calling

• output directory structure improved to reduce nesting, as a result of related changes to outputs

• module names use consistent underscore separation, more descriptive

• GAM filtering: added a param "--gamFilterMore". When false, filters only on primary alignment score + defrays ends. When true, applies MAPQ filter and discards unmapped reads

• Under the hood this is controlled by conditional logic in modules.config that sets the value of ext.args2

• Fastp previously defaulted to discard unmerged reads for aDNA samples. It now defaults to keeping them.

• This is controlled by a parameter --discardUnmerged (default false)

• Under the hood this controls FASTP ext.args in modules.config, outputting two extra fastq files for unmerged reads.

• If these files are present, they are first concatenated with the respective merged reads before sample level concatenation and re-deduplication

• i.e. the mapper would receiving one file per sample, containing all libraries, and both merged and unmerged reads in a single deduplicated package

• As a result of this change, also added safety to FASTQC on receiving empty FASTQ files

• Patched non-zero exit when getting heap space for profilepmd, which would cause process to fail even if successful. Force 0 exit.

• All output folders use underscores (no dashes) for consistency

• updated DAG

• additional test graphs added to data/test/additional_graphs

• added safety to FASTQC in case of empty input files (e.g. in the case of unmerged aDNA output from FASTP)

1.2.2

Patched profiles and added cluster resource limits

1.2.1

Fixed

• Patched profilepmd heap space issues, dynamically get maximum heap space and provide to profilepmd as upper limit

1.2.0

Changed

• Migrated to codespaces from gitpod

1.1.0

Added

• pixi environments, accounting for systems with custom Apptainer installations (e.g. systems where user namespace is not allowed). default lacks apptainer and will look for system-wide Apptainer, apptainer provides Apptainer support in the pixi environment

• option to skip profile pmd process

Changed

• Batch submission shell script accounts for older tmux versions & other minor improvements

• Safety added to projectDir cleanup shell script

Fixed

• addressed java heap space for PROFILEPMD

1.0.0

[1.0.0] - 2025-02-19

Added

• Initial grave release

Dependencies

• pixi