Create nf-core Pipeline

Generate a new Nextflow pipeline using the nf-core base template following community best practices.

Quick Commands

bash

# Interactive pipeline creation (recommended for first time)
conda run -n nf-core nf-core pipelines create

# Create with all options specified
conda run -n nf-core nf-core pipelines create \
    --name mypipeline \
    --description "My pipeline description" \
    --author "Your Name"

# Create using a template YAML file
conda run -n nf-core nf-core pipelines create --template-yaml template.yml

# Create with custom organization prefix
conda run -n nf-core nf-core pipelines create --org myorg

Process

•Validate Pipeline Name: Must be lowercase, no punctuation, descriptive
•Run Create Command: Execute the interactive wizard or provide flags
•Select Template Features: Choose which features to include/exclude
•Initialize Git: Automatic git init with proper branches
•Push to GitHub: Create empty repo and push initial commit

Pipeline Naming Rules

•Lowercase only: rnaseq not RNAseq
•No punctuation: methylseq not methyl-seq
•Descriptive: Name should indicate data type or analysis
•Unique: Check existing pipelines at https://nf-co.re/pipelines

Template Features (Selectable)

During creation, you can include/exclude:

Feature	Description	Default
`github`	GitHub integration, CI tests, templates	Yes
`ci`	Continuous integration with linting	Yes
`github_badges`	README badges (CI status, DOI, etc.)	Yes
`igenomes`	iGenomes reference genome config	Yes
`nf_core_configs`	nf-core institutional configs	Yes
`code_linters`	Pre-commit hooks, Prettier	Yes
`citations`	CITATIONS.md file	Yes
`gitpod`	Gitpod development environment	No
`codespaces`	GitHub Codespaces support	No
`multiqc`	MultiQC report integration	Yes
`fastqc`	FastQC quality control module	Yes
`nf_schema`	nf-schema parameter validation	Yes
`slackreport`	Slack notifications	No
`adaptivecard`	Microsoft Teams notifications	No

Template YAML File

Create template.yml for reproducible pipeline creation:

yaml

name: mypipeline
description: "A pipeline for analyzing XYZ data"
author: "Your Name"
org: nf-core  # or your organization

# Skip specific features
skip:
  - igenomes
  - slackreport
  - adaptivecard
  - gitpod

# Or explicitly include features
# features:
#   - github
#   - ci
#   - multiqc

Generated Structure

code

nf-core-mypipeline/
├── .github/                  # GitHub Actions, templates
├── .devcontainer/            # Dev container config
├── assets/                   # Email templates, logos
├── bin/                      # Custom scripts
├── conf/                     # Configuration files
│   ├── base.config          # Base resource config
│   ├── modules.config       # Module-specific config
│   └── test.config          # Test profile config
├── docs/                     # Documentation
│   ├── output.md            # Output documentation
│   └── usage.md             # Usage documentation
├── lib/                      # Groovy libraries
├── modules/                  # nf-core modules
│   ├── local/               # Local modules
│   └── nf-core/             # Installed modules
├── subworkflows/            # Subworkflows
│   ├── local/               # Local subworkflows
│   └── nf-core/             # Installed subworkflows
├── workflows/               # Main workflow
├── main.nf                  # Entry point
├── nextflow.config          # Main config
├── nextflow_schema.json     # Parameter schema
├── CHANGELOG.md             # Version history
├── CITATIONS.md             # Tool citations
├── LICENSE                  # MIT License
└── README.md                # Pipeline README

Git Branches

The template sets up three branches:

•master / main: Stable releases only
•dev: Active development
•TEMPLATE: Vanilla template for sync

After Creation

•

Create GitHub Repository:

bash

# Create empty repo on GitHub (no README, no license)
# Then push:
cd nf-core-mypipeline
git remote add origin git@github.com:username/nf-core-mypipeline.git
git push -u origin --all

•

Install Modules:

bash

conda run -n nf-core nf-core modules install fastqc
conda run -n nf-core nf-core modules install multiqc

•

Run Lint:

bash

conda run -n nf-core nf-core pipelines lint

•
Test Pipeline:
bash
```
nextflow run . -profile test,docker
```

Important Notes

•Always discuss new pipeline proposals on nf-core Slack before development
•Check if similar pipeline exists: https://nf-co.re/pipelines
•Development starts on dev branch, not master
•PRs target dev branch for nf-core pipelines