2.0.7

• Fixed issue with summarize_alignments, compress_alignments and create_bar_chart tasks being re-run when resuming from successfully completed run.

• Fixed error running split_alignments_by_sample.R when an input dataset has no sampled reads, either because the input data has no reads of the sampling returns no reads due to filtering based on trimming; now uses the sample sheet to know what samples are present and creates an output file for each regardless of whether that sample was contained in the chunk.

2.0.6

• Created inner workflow containing much of the workflow logic and that can be incorporated as a sub-workflow within another data processing pipeline that already has a channel for FASTQ files and sample metadata stored in another format; updated main workflow by removing the logic and calling the inner workflow instead

• Moved the process definitions from mga2.nf to a separate file, processes.nf

• New add_sample_ids process for adding a numeric id to the sample sheet which is used instead of the user-specified id within the workflow logic

• added checks for missing bowtie index for matched genomes in check_inputs process

• Updated Docker container with more recent versions of Rust, R and dependent R packages; use of separate build directory to avoid R scripts in the container taking precedence on the path over the versions in the bin directory

2.0.5

• Reworked alignment summarization R script so that this is done for each sample separately (more efficient for large numbers of samples), new processs added for splitting and collating chunked alignment output files

• Added process for compressing the final alignment files

• Tasks for trimming and splitting input FASTQ sequences and splitting alignments no longer use the local executor but run as separate jobs

• Added run option for Singularity to prevent it mounting the users home directory, removed cacheDir setting for the Singularity image so will now be built in a work subdirectory

• Docker container now based on Debian slim base image, updates to more recent versions of R and dependent R packages, added rust to the conda environment

• Now capable of handling of large bowtie indexes with ebwtl suffix

2.0.4

• Reworked summarize_alignments.R to read in alignments for all datasets into single data frame instead of splitting into separate chunks following occasional hanging jobs on CRUK CI HPC cluster running read_tsv_chunked operation

• Changed exit status codes used for Nextflow error/retry strategy

2.0.3

• Updated Conda environment (Miniconda3 py39_4.12.0) and added check for SHA-256 checksum

• Updated R to 4.1.3 (tidyverse 1.3.1 packages in conda-forge were built with this version of R, were being flagged in warnings)

2.0.2

• Updated Conda environment (Miniconda3 py39_4.11.0) and dependencies (R 4.1.3, tidyverse 1.3.1, bowtie 1.3.1)

• Change build and deployment directory to /opt/mga2 and add to PATH for ease of use in a parent pipeline that uses MGA as a sub-workflow.

• Fixed issue with creating summary plot (create_bar_chart.R) when there are no reads in any of the input FASTQ files (caused the category variable to be treated as a logical and resulted in failed replace_na call)

• Improved handling of required columns and column types in R scripts

2.0.1

• Added version tag for container in nextflow.config

2.0.0

Initial release of a rewrite of the MGA
contaminant screen for high-throughput sequencing data.

• workflow ported to Nextflow
• sampling of FASTQ records rewritten in Rust (this is the rate limiting step for very large data sets such as those from NovaSeq S4 flow cells)
• summarization and plotting rewritten using R and the dplyr and ggplot2 packages from the tidyverse

2.0.0-alpha

Updated version for 2.0.0 release