nf-core/rnaseq 3.28.0

nf-core/rnaseq v3.28.0 - Nickel Narwhal

on GitHub

A release with new options to fail a run on failed per-sample checks, a move to the nf-core-utils plugin, and several bug fixes. Two changes alter behaviour and are marked below.

New features

• New --fail_on_low_trimmed_reads, --fail_on_low_mapped_reads and --fail_on_undetermined_strandedness options fail the run instead of silently skipping the sample or defaulting to unstranded (#1969)

Fixes

• Fix HISAT2_ALIGN running for only one sample when the HISAT2 index is built in-run (#1994 by @janaki-bio)
• Apply --min_mapped_reads filtering to HISAT2 alignments (#1943)
• Leave samples failing --min_mapped_reads out of Salmon/RSEM quantification and the merged count tables, so such samples no longer get aligner-based counts (behaviour change), and document the MultiQC fail tables (#1984)
• Stop building the SortMeRNA index twice, and build the Salmon strandedness-inference index centrally in PREPARE_GENOME_INDICES (#1958)
• Skip the strandedness-inference Salmon index for BAM-only auto samples, stop an unparsed HISAT2 alignment rate dropping every sample, and fix the EAUTILS_GTF2BED conda lock paths (#2001)
• Fix --with_umi transcriptome-side samtools stats silently colliding with genome-side stats (#1942)
• Stop a library with no quantified reads crashing the SummarizedExperiment merge and DESeq2 QC (#1986)
• Fix duplicated valueless flags passed via --extra_trimgalore_args (#1968 by @jaredfischbach)
• Remove the cpus = 1 override for RiboDetector (#1972 by @jaredfischbach)
• Right-size RSEM_CALCULATEEXPRESSION resources, drop its unused BAM output, and stop pinning --seed 1 in its default arguments (#1959)
• Use size() instead of the size property on lists in samplesheet validation, which fails on Nextflow 26.11 and later (#1990 by @bentsherman)
• Declare force_completion_email in the config and schema, and test --fail_on_low_trimmed_reads and --fail_on_low_mapped_reads (#1999)

Updates

• Replace the vendored utils_* subworkflows with the nf-core-utils@0.7.0 plugin; the completion email is skipped on Seqera Platform launches unless params.force_completion_email = true (behaviour change) (#1977 by @edmundmiller)
• Update Salmon to 2.11.0, restoring --noLengthCorrection for 3' tag data such as QuantSeq (#1991)
• Update Parabricks to 4.8.0-1, so the Parabricks run log no longer overwrites STAR's Log.final.out (#1978 by @gburnett-nvidia)
• Template updates for nf-core/tools v4.2.1 and v4.2.2 (#1980, #2000)
• Module bumps: FastQC 0.13.0, fastp 1.4.0, samtools 1.25, and bowtie2 2.5.5 with fixed version parsing and rebuilt containers (#1993, #1997)

See the full changelog for all 25 merged PRs.

Credits

Special thanks to the following for their contributions to the release:

• Adam Talbot
• Ben Sherman
• Brandon Seah
• Chris Fields
• Daniel Lundin
• Dominik Maszczyk
• Edmund Miller
• Emmanuel Tan
• Felix Krueger
• Friederike Hanssen
• Gary Burnett
• Janaki B
• Jared R Fischbach
• Jenny Drnevich
• Jose Espinosa-Carrasco
• Louis Le Nézet
• Matthias Hörtenhuber
• Maxime U Garcia
• NickPanyushev
• Paolo Di Tommaso
• Paolo Inglese
• Phil Ewels
• Rayan Hassaïne
• Scott Handley
• Sriiraam
• Sushant Hona

Full Changelog: 3.27.0...3.28.0