nf-core pipeline · fusion, fusion-genes, gene-fusion
nf-core/rnafusion
A Nextflow bioinformatics pipeline for detecting and visualising gene fusions and splicing events from RNA-seq data.
What nf-core/rnafusion is
nf-core/rnafusion is an automated bioinformatics analysis pipeline designed to discover, quantify, and visualise gene fusions—abnormal hybrid genes formed by the structural joining of two separate genes—from RNA sequencing data. Developed by Martin Proks, Maxime Garcia, and Annick Renevey within the nf-core community, it provides a standardised framework for identifying chimeric transcripts across multiple analytical engines.
Because individual fusion-calling algorithms differ in their sensitivity and error profiles, combining multiple tools is standard practice. The pipeline analyses raw sequence data across several callers and consolidates the predictions into shared reports and visualisation documents, helping researchers evaluate and prioritise candidate fusion events alongside splicing variations.
What the pipeline does
The workflow begins by verifying the input samplesheet and assessing raw read quality with FastQC, with optional adapter and quality trimming handled by fastp. Reads are aligned to the reference genome using STAR. Fusion transcript detection is then executed across three independent tools: Arriba, STAR-Fusion, and FusionCatcher (with optional 3-prime trimming). In parallel, StringTie performs transcript assembly and quantification, while CTAT-SPLICING identifies cancer-associated splicing aberrations.
Identified fusion candidates are integrated and annotated using Fusion-report. The workflow performs post-processing inspections with FusionInspector, compiles calls into a Variant Call Format (VCF) file, produces Arriba visualisation graphics, and derives sequencing quality metrics through Picard and GATK MarkDuplicates. Finally, SAM/BAM alignments are converted to CRAM format with samtools, and MultiQC aggregates all sample-level quality control outputs into an HTML report.
Key facts
- Pipeline engine
- Nextflow
- Supported genome build
- GRCh38 only
- Fusion detection callers
- Arriba, STAR-Fusion, FusionCatcher
- Supported runtime environments
- Docker, Singularity (Conda is not supported)
- Primary outputs
- VCF, PDF visualisations, HTML and TSV reports, CRAM alignments, MultiQC report
- Transcriptome analysis tools
- StringTie, CTAT-SPLICING
- Latest release
- 4.1.3
What you can use it for
- Oncogenic fusion detection
- Identifying gene fusions and chimeric transcripts from tumour RNA sequencing data using three complementary detection callers.
- Integrated fusion reporting
- Consolidating predictions from Arriba, STAR-Fusion, and FusionCatcher into merged VCF, TSV, and HTML reports via Fusion-report.
- Cancer splicing analysis
- Screening for aberrant splicing events with CTAT-SPLICING alongside transcript assembly and quantification using StringTie.
- Visual verification of fusion breakpoints
- Generating graphical outputs and inspection models using Arriba visualisation and FusionInspector to review read support for candidate fusions.
Run it here
The research agent runs nf-core/rnafusion with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
Strengths and limits
Good at
- Runs Arriba, STAR-Fusion, and FusionCatcher concurrently and consolidates results using Fusion-report.
- Provides integrated visualisations including dedicated PDF diagrams and FusionInspector analyses.
- Pairs fusion detection with cancer splicing aberration screening and StringTie transcript quantification.
- Validated on release using automated continuous integration runs on full-sized datasets via AWS infrastructure.
Watch out for
- GRCh38 is the only supported reference genome assembly.
- Conda is not supported, requiring the use of Singularity or Docker containers.
- References for Arriba and FusionCatcher cannot be generated automatically and must be downloaded from external sources or S3.
- FusionCatcher cannot be used on single-end reads that are shorter than 130 base pairs.
- Building references locally is computationally demanding, taking more than 24 hours on high-performance computing clusters.
Questions
Can I run nf-core/rnafusion without my own cluster?
The research agent runs nf-core/rnafusion with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
What licence is nf-core/rnafusion under?
MIT, like every nf-core pipeline.
Which reference genome assemblies does the pipeline support?
The pipeline only supports the human GRCh38 reference genome.
Can single-end RNA sequencing reads be used?
Paired-end reads are recommended, with single-end reads considered a last resort. FusionCatcher cannot process single-end reads that are shorter than 130 base pairs.
Is Conda supported for running this pipeline?
No, Conda is not supported. The pipeline must be run using container runtimes such as Singularity or Docker.
How are reference genomes and tool databases provided?
Pre-built references can be downloaded from the nf-core AWS S3 megatests bucket. While the pipeline can generate reference indices for Salmon, STAR, and STAR-Fusion, references for Arriba and FusionCatcher cannot be generated by the workflow and must be downloaded.
How to cite it
Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.
Written from nf-core/rnafusion's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.