nf-core/rnaseq 3.23.0 (Gallium Gecko) is out! 🦎 🎉 This is a big one - highlights: 🦠 Prokaryotic RNA-seq support New

-profile prokaryotic

configures the pipeline for bacterial/archaeal data: Bowtie2+Salmon alignment, GFFRead transcript extraction for CDS-only annotations, and automatic STAR CDS configuration. This isn't a specialist prokaryotic pipeline, so the support is basic, but it's been much asked-for! Thanks to @Juliana Assis, @Sebastian Schulz, @Albert Palleja Caro and @Marine Cambon for recommendations and testing, I think this will be a subject of revision going forward! 🏎️ GPU-accelerated alignment NVIDIA Parabricks

rna_fq2bam

support for GPU-accelerated STAR alignment and mark duplicates (

--use_parabricks_star

). Thanks @Gary Burnett! 🧬 Contamination screening with Sylph New

--sylph_db

and

--sylph_taxonomy

params for fast, lightweight contamination detection. Thanks @Ezra Greenberg! 🧹 Multiple rRNA removal tools Choose between SortMeRNA (default), Bowtie2, or RiboDetector via

--ribo_removal_tool

. 📊 Improved RSEM quantification tximport now processes RSEM outputs, producing length-scaled counts, gene/transcript length matrices, and SummarizedExperiment objects - matching the Salmon/Kallisto pathway. Previous RSEM merge script outputs are preserved in

rsem_merge_counts/

. ✨ Other notable changes: - 🔍 FastQC step after BBSplit/rRNA filtering for QC on reads actually used for alignment - 📂 Uncompressed FASTQ input support - 🏷️ UMI dedup now works with

--aligner star_rsem

- 🔖 Per-sample

seq_platform

and

seq_center

samplesheet columns for BAM read group tags - ⚡ Pseudo-alignment with pre-built index no longer requires

--fasta

- 🔗 Single-library samples now included in merged fastq output - ⚠️ Informative error when Salmon strandedness inference fails - ✈️ Offline mode fix for

igenomes_base

validation - ✅ Nextflow 25 strict syntax compatibility - 📦 Topic-based version reporting for modules Thanks to everyone who contributed to this release: @Albert Palleja Caro, @chase mateusiak, @Christian Mertes, @Elad Herz, @Ezra Greenberg, @Gary Burnett, @Isaac Virshup, @Juliana Assis, @Marine Cambon, @Matthias Zepper, @Sebastian Schulz - and to @Adam Talbot, @Louis Le Nézet, and @Maxime Garcia for thorough reviews on the release PR. ❤️❤️ Full changelog and release: https://github.com/nf-core/rnaseq/releases/tag/3.23.0 As always, bugs and feature requests to https://github.com/nf-core/rnaseq/issues!

Oh wow, that release reads like a 4.0.0 actually - features galore! Thanks Jonathan for your hard work! thanks

I did think about it, but there are no breaking changes really, so no semver grounds for a 4.0 😄

Next time 😉

🤔 So we will need to wait for DSL3 or rewrite the pipeline in Snakemake snakemake to see the 4.0.0 happen?

Next time 😉

Next time that there will be so many new additions? You obviously still have big plans for this pipeline…

You'll get your 4.0 when we break the interface 😄

Thank you so much for the update and all the work! That's a work for this week to run it here 😃

You may well find issues to file, we can do a patch release later 🙂

Hello. The sample output on the nf-co.re/rnaseq website for star-rsem is missing the merged rsem files and my pipeline run also. Do you have recommendations to recover the counts?

The infrastructure that populates the outputs on the website has been variously broken for a while and I can see that the run failed for non-workflow related reasons. Were there errors in your pipeline run? We did slightly change how RSEM outputs look to introduce consistency with salmon - see https://nf-co.re/rnaseq/3.23.0/docs/output/#star-via-rsem. But if you're not seeing that documented structure you probably have a pipeline failure of some kind.

@Jonathan Manning thank you for sharing your experiences. We ran the pipeline on AWS Batch as the dataset sizes did not allow for local runs. We didn't find any errors in the logs. In the end thankfully we found the merged RSEM files (gene, transcript, counts and TPMs) in the work directory.