nextflow

Data-driven workflow orchestration for scalable scientific pipelines

TLDR

Run pipeline

$ nextflow run [pipeline.nf]

Run with parameters

$ nextflow run [pipeline.nf] --input [data.txt]

Run pipeline from GitHub

$ nextflow run [user/repo]

Resume failed run from checkpoint

$ nextflow run [pipeline.nf] -resume

Run with Docker containers

$ nextflow run [pipeline.nf] -with-docker

Run with a specific config profile

$ nextflow run [pipeline.nf] -profile [docker]

Generate execution report and timeline

$ nextflow run [pipeline.nf] -with-report -with-timeline

Show pipeline info

$ nextflow info [pipeline]

List previous runs

$ nextflow log

Clean up cached work directories

$ nextflow clean -f

SYNOPSIS

nextflow [options] command [arguments]

nextflow is a reactive workflow framework and DSL for writing data-intensive computational pipelines. It enables the rapid composition of parallel and distributed pipelines using isolated processes connected by dataflow channels.Nextflow abstracts execution from pipeline logic, allowing the same workflow to run locally, on HPC clusters (SLURM, SGE, PBS, LSF), or on cloud platforms (AWS Batch, Azure Batch, Google Cloud, Kubernetes). It provides built-in support for containers (Docker, Singularity, Podman) and package managers (Conda), and uses content-addressable caching to enable safe resumption of failed runs.

PARAMETERS

run PIPELINE

Execute a pipeline project (local path, script, or GitHub repo).

pull PROJECT

Download or update a remote project.

info [PROJECT]

Show system or pipeline info.

log

Show history of executions.

clean

Clean up work directories from past executions.

drop PROJECT

Delete a local project cache.

list

List downloaded pipeline projects.

-resume [SESSION]

Resume execution from cached results; optionally specify session ID.

-with-docker [IMAGE]

Enable Docker container execution.

-with-podman [IMAGE]

Enable Podman container execution.

-with-singularity [IMAGE]

Enable Singularity container execution.

-with-conda [ENV]

Enable Conda environment management.

-profile NAME

Select a configuration profile defined in nextflow.config.

-c FILE, -config FILE

Use an alternative configuration file.

-params-file FILE

Load pipeline parameters from JSON/YAML file.

-w DIR, -work-dir DIR

Directory for intermediate task files (default: ./work).

-with-report [FILE]

Generate an HTML execution report.

-with-timeline [FILE]

Generate a timeline HTML file.

-with-trace [FILE]

Generate an execution trace file.

-with-dag [FILE]

Generate a workflow DAG diagram.

-bg

Run in background.

-r REV

Use a specific Git revision/branch/tag.

-h, -help

Display help information.

CAVEATS

Requires Java 11 or later. Workflow scripts use a Groovy-based DSL (DSL2 is the current standard). The work directory can grow large across runs; use `nextflow clean` to reclaim space. Pipelines pulled from GitHub are cached in `~/.nextflow/assets/`.

HISTORY

Nextflow was created by Paolo Di Tommaso at the Centre for Genomic Regulation (CRG) in Barcelona, with initial public release in 2013. It was designed for reproducible, scalable scientific workflows, initially in bioinformatics. The project is now maintained by Seqera Labs and underpins the nf-core community of curated pipelines.