Under active development

LIANA-NicheNet Pipeline

A Nextflow pipeline for integrated cell-cell communication analysis using LIANA (Python) and NicheNet (R).

Quick Start

You can run this pipeline directly from GitHub with Nextflow:

```bash

Run with test data (requires Nextflow and Conda)

nextflow run yourusername/liana-nichenet-pipeline -r main \ --input yourdata.h5ad \ --sendercelltype TypeA \ --receiver_celltype TypeB \ -profile conda

Run on an HPC with SLURM

nextflow run yourusername/liana-nichenet-pipeline -r main \ --input yourdata.h5ad \ --sendercelltype TypeA \ --receiver_celltype TypeB \ -profile slurm,singularity ```

Pipeline Overview

This pipeline integrates:

1. LIANA: For ligand-receptor interaction identification
2. NicheNet: For downstream gene regulation prediction

The workflow: - Runs LIANA on AnnData object to prioritize ligand-receptor pairs - Feeds these pairs into NicheNet for downstream analysis - Creates integrated visualizations connecting cell communication to gene regulation

Requirements

• Nextflow (≥21.10.3)
• One of the following:
  • Docker or Singularity (recommended for HPC)
  • Conda/Mamba

Input Requirements

• An AnnData object (.h5ad file) with scRNA-seq data
• Cell type annotations should be in adata.obs['cell_type']
• Sender and receiver cell type names (must match names in the data)

Parameters

| Parameter | Description | Required | Default | |-----------|-------------|----------|---------| | --input | Path to input .h5ad file | Yes | None | | --sender_celltype | Name of sender cell type | Yes | None | | --receiver_celltype | Name of receiver cell type | Yes | None | | --outdir | Output directory | No | results |

Output

The pipeline produces:

results/ ├── liana/ │ ├── liana_ranked_interactions.csv # Ranked L-R interactions │ ├── top_lr_pairs_for_nichenet.csv # Top pairs for NicheNet input │ └── liana_dotplot.pdf # Visualization ├── nichenet/ │ ├── nichenet_results_from_liana.rds # R object with results │ ├── ligand_activity_heatmap.pdf # Activity predictions │ └── ligand_target_heatmap.pdf # Target gene regulation └── integrated_results/ ├── liana_nichenet_integrated_results.csv # Combined results ├── liana_nichenet_correlation.pdf # Method correlation └── top_ligands_summary_heatmap.pdf # Ranking comparison

Execution Profiles

| Profile | Description | |---------|-------------| | conda | Uses conda environments defined in ./envs/ | | docker | Uses Docker containers | | singularity | Uses Singularity containers (recommended for HPC) | | slurm | Configuration for SLURM scheduler | | sge | Configuration for SGE scheduler |

Combine profiles with comma: -profile slurm,singularity

Citation

If you use this pipeline, please cite:

• LIANA: https://github.com/saezlab/liana-py
• NicheNet: https://github.com/saeyslab/nichenetr
• Nextflow: Di Tommaso, P., et al. (2017). Nextflow enables reproducible computational workflows. Nature Biotechnology, 35(4), 316-319.