RNA seq STAR Salmon Example

The built-in Demo 3 - nf-core/rnaseq rebuilt from nodes example rebuilds the default route of nf-core/rnaseq from library nodes. Every step is a node you can open, inspect (its Code tab shows the Nextflow it adds) and change. It is also the reference for building your own pipelines in N-WAVE: nothing in it is special-cased in the app.

What's on the canvas

Part	Nodes
Inputs	Samplesheet (4 samples of the nf-core/rnaseq test data: 3 paired-end, 1 single-end) and Reference genome (a Parameters node with the test genome FASTA and gzipped GTF).
Genome preparation	Unzip GTF (nf-core/gunzip) → Filter GTF (nf-core/custom/gtffilter, which drops annotation on sequences that aren't in the genome) → Transcript FASTA (nf-core/gffread -w), STAR index (nf-core/star/genomegenerate) and Salmon index (nf-core/salmon/index).
Reads	FastQC (raw reads) and Trim Galore.
Alignment and quantification	STAR alignment (nf-core/star/align, with the options nf-core/rnaseq 3.27.0 uses, including --quantMode TranscriptomeSAM) → Salmon quantification (nf-core/salmon/quant, alignment mode, --libType A).
Statistics	SAMtools sort, index, stats (the nf-core/bam_sort_stats_samtools subworkflow).
Report	Collect QC reports (a channel operator that mixes every report into one channel) → MultiQC.

Notes on the canvas explain each part. Every node uses only library features: settings like STAR's options are the node's Extra arguments, and output names are its Output prefix.

Opening it

Open Demo 3 - nf-core/rnaseq rebuilt from nodes on the home page. Its 11 nf-core components (10 modules and one subworkflow) aren't installed on a fresh install. A banner lists them; click Install them. The subworkflow brings in five samtools modules and the bam_stats_samtools subworkflow. In the browser demo they install into the browser.

The example is read-only. Opening panels and Code tabs doesn't change it. Changing a setting or moving a node makes an editable copy.

Running it (Docker install)

Run Workflow in the bottom bar. The run takes a few minutes on a laptop; the first run also pulls the module containers.

Each nf-core node saves its outputs to results/<node name>/:

• results/salmon_quantification/<sample>/quant.sf: transcript quantifications.
• results/star_alignment/: alignments and STAR logs.
• results/samtools_sort_index_stats/: sorted BAMs, indexes and statistics.
• results/multiqc/multiqc_report.html: the report, covering FastQC, Cutadapt (Trim Galore), STAR, Salmon and SAMtools for every sample.

When the run finishes, the run panel lists Results. Click multiqc/multiqc_report.html to open the report, or expand All result files.

Running the exported project

Export → Download project downloads main.nf, nextflow.config, the nf-core modules and subworkflows it uses, inputs/samplesheet.csv and a README. With Nextflow and Docker installed:

nextflow run main.nf -profile docker

The results go to results/, in the same layout as above.

Manual check (release testing)

1. Start a fresh Docker install of N-WAVE (docker compose up -d).
2. Open Demo 3 - nf-core/rnaseq rebuilt from nodes, click Install them and wait for the banner to go.
3. Click Run Workflow. The run should finish with "Workflow Complete".
4. In the run panel's Results, open multiqc/multiqc_report.html. It should list four samples with FastQC, Cutadapt, STAR, Salmon and SAMtools sections.
5. In All result files, check that salmon_quantification/<sample>/quant.sf exists for all four samples.
6. Click Export → Download project, unzip it, run nextflow run main.nf -profile docker and check that results/multiqc/multiqc_report.html and the four quant.sf files are there.

Using your own data

1. Duplicate the example.
2. Replace the samplesheet rows and the reference FASTA and GTF (an uncompressed GTF works too: delete Unzip GTF and connect the GTF to Filter GTF).
3. In STAR index, remove --genomeSAindexNbases 7 from the extra arguments. It only suits the tiny test genome.
4. Raise CPUs and memory in Execution Settings → Resources: a human or mouse STAR index needs about 40 GB of memory.

What the full pipeline does in addition

nf-core/rnaseq also:

• merges several runs of a sample;
• infers strandedness;
• removes rRNA (SortMeRNA, RiboDetector) and contaminant reads (BBSplit);
• offers other aligners and quantifiers (HISAT2, RSEM, Bowtie 2, Kallisto) and UMI handling;
• adds QC (RSeQC, Qualimap, dupRadar, Preseq, DESeq2 PCA);
• builds gene-level count matrices with tximport.

To run all of that, use Demo 2 - nf-core/rnaseq pipeline, which runs the whole pipeline as one node (Running the nf-core/rnaseq example).

The nodes it uses are described in the Node reference and the nf-core library.