nextflow

Nextflow: Workflow Management

Overview

Nextflow is a powerful workflow management system designed for computational biology and bioinformatics. It enables the creation of portable, reproducible pipelines that can run on local machines, HPC clusters, and cloud platforms.

When to Use This Skill

This skill should be used when:

• Creating scalable bioinformatics pipelines
• Running workflows on HPC or cloud
• Building nf-core pipelines
• Ensuring reproducibility
• Managing containerized workflows

Quick Start

Installation

# Using conda
conda install -c bioconda nextflow

# Direct download
curl -s https://get.nextflow.io | bash

# Or using sdkman
sdk install nextflow

Basic Pipeline

// main.nf
process SAYHELLO {
    input:
        val x

    output:
        stdout

    script:
    """
    echo "Hello $x!"
    """
}

workflow {
    Channel.from('World', 'Nextflow', 'Pipeline') \
        | SAYHELLO \
        | view
}

Single-Cell Example

Basic scRNA-seq Pipeline

// main.nf
paramsreads = "data/reads/*_{1,2}.fastq.gz"
paramsgenome = "genome.fa"
paramsindex = "genome.index"

process FASTQC {
    input:
    path reads

    output:
    path "fastqc_reports"

    script:
    """
    mkdir -p fastqc_reports
    fastqc $reads -o fastqc_reports
    """
}

process ALIGN {
    input:
    path reads
    path index

    output:
    path "aligned.bam"

    script:
    """
    hisat2 -x $index -1 ${reads[0]} -2 ${reads[1]} -S aligned.sam
    samtools view -b aligned.sam > aligned.bam
    """
}

process COUNT {
    input:
    path bam
    path gtf

    output:
    path "counts.txt"

    script:
    """
    featureCounts -a $gtf -o counts.txt $bam
    """
}

workflow {
    reads = Channel.fromFilePairs(paramsreads)
    FASTQC(reads)
    ALIGN(reads, paramsindex)
    COUNT(ALIGN.out, params.gtf)
}

Running Nextflow

Execution Modes

# Local
nextflow run main.nf

# Docker
nextflow run main.nf -dockerize

# Singularity
nextflow run main.nf -with-singularity

# Conda
nextflow run main.nf -with-conda