nf-core Pipelines Skills

nf-core is a community-driven collection of 143+ curated Nextflow pipelines for bioinformatics. All pipelines are open-source (MIT), rigorously tested, and run portably on laptops, HPCs, and cloud platforms with automated dependency management via Docker, Singularity, or Conda.

Available Skills

Getting Started & Usage

Installation, configuration, and common usage patterns for running any nf-core pipeline on local machines, HPC clusters, or cloud environments.

Skill file: nfcore_usage.md

When to use:

• First time setting up Nextflow and nf-core
• Configuring pipelines for your HPC cluster or cloud environment
• Understanding resource management, resume, and offline execution
• Looking up nf-core CLI tool commands

Single-Cell & Bulk RNA-seq Pipelines

Pipelines for processing single-cell RNA-seq (10x, Drop-seq, Smart-seq) and bulk RNA-seq data from raw FASTQs to count matrices.

Skill file: nfcore_transcriptomics.md

When to use:

• Processing 10x Chromium, Drop-seq, or Smart-seq scRNA-seq data
• Running downstream single-cell analysis (doublet removal, integration, annotation)
• Processing bulk RNA-seq with STAR, HISAT2, Salmon, or Kallisto
• Generating gene/transcript count matrices and QC reports

Spatial Omics Pipelines

Pipelines for spatial transcriptomics platforms including Visium, Xenium, MERSCOPE, CosMX, and molecular cartography.

Skill file: nfcore_spatial.md

When to use:

• Processing 10x Visium or Visium HD data
• Analyzing Xenium in situ data with cell segmentation
• Running technology-agnostic spatial pipelines (sopa)
• Processing Resolve Bioscience Molecular Cartography data

Epigenomics Pipelines

Pipelines for chromatin accessibility, histone modification, protein-DNA interaction, and DNA methylation profiling.

Skill file: nfcore_epigenomics.md

When to use:

• Processing ATAC-seq data (bulk)
• Analyzing ChIP-seq experiments with peak calling
• Running CUT&Run or CUT&Tag with spike-in normalization
• Processing bisulfite sequencing or TAPS methylation data

Variant Calling Pipeline (Sarek)

Germline and somatic variant detection from WGS, WES, or targeted sequencing data with 16+ variant callers.

Skill file: nfcore_variant_calling.md

When to use:

• Detecting germline or somatic SNVs, indels, SVs, and CNVs
• Processing tumor/normal pairs or tumor-only samples
• Running multi-caller consensus variant analysis
• Annotating variants with SnpEff or VEP

Hi-C Chromatin Conformation Pipeline

Pipeline for processing Hi-C chromosome conformation capture data to study 3D genome organization: contact maps, TADs, and A/B compartments.

Skill file: nfcore_hic.md

When to use:

• Processing Hi-C data (digestion or DNase protocol)
• Generating multi-resolution contact maps (.cool/.mcool)
• Calling TADs and A/B compartments
• Studying 3D genome organization and chromatin interactions

Dynamic Pipeline Discovery (All 143+ Pipelines)

Meta-skill for dynamically discovering and using any nf-core pipeline, including those not covered by the skill files above. Teaches the agent how to fetch pipeline documentation, parameters, and samplesheet formats on-the-fly from standardized nf-core URLs and schemas.

Skill file: nfcore_dynamic_discovery.md

When to use:

• User asks about a pipeline not covered in the detailed skill files above
• Exploring what pipelines are available for a specific data type
• Need to look up parameters or samplesheet format for any nf-core pipeline
• Pipeline has been updated and you need the latest information

[!TIP] The detailed skill files above cover the most commonly used pipelines with full parameter tables and examples. For all other pipelines, use the dynamic discovery skill to fetch information on-the-fly from nf-co.re.

Using Skills

1. Start with usage guide: Read nfcore_usage.md for installation and configuration
2. Select pipeline skill: Choose the skill matching your data type
3. Pipeline not listed? Use nfcore_dynamic_discovery.md to fetch docs on-the-fly
4. Follow samplesheet format: Each pipeline requires a specific CSV samplesheet
5. Test first: Always run with -profile test,docker before real data
6. Use -resume: Re-run failed pipelines without recomputing successful steps