nf-pseudobulk

nf-pseudobulk is a Nextflow pipeline used to perform Gene Set Enrichment Analysis (GSEA) on pseudobulk data.

• Runs pseudobulk aggregation on scRNA-seq h5ad files by summing expresesion values per patient and per cell type
• Runs GSEA for the sum of cell types and for each resulting pseudobulk sample

h5ad Preprocessing

The h5ad file for use as input in this workflow should have the following characteristics:

• Counts data, with the preprocessing of your choosing, stored as a layer. The name of this layer should be provided in the input samplesheet (more details below). Please note: conversion to counts values to z-scores is not advised, as subsequent steps in the processing don't allow negative values.
• Genes identified by their Gene Symbol.

Usage

Before executing the workflow, create a Nextflow secret called SYNAPSE_AUTH_TOKEN using a Synapse Personal Access Token.

To run the pipeline with docker use:

nextflow run CRI-iAtlas/nf-pseudobulk --input <path/to/input.csv> -profile docker

Input Samplesheet

The input to this pipeline is a CSV samplesheet specified with the --input parameter

An example input sheet can be found at data/test_samplesheet.csv

Samplesheet requirements:

• dataset: Name of dataset
• h5ad: Synapse ID of input h5ad file to process
• counts_layer: Name of layer in h5ad with raw counts (default: counts)
• sample_id: Name of column in h5ad containing Sample IDs
• cell_type_id: Name of column in h5ad containing Cell Type ID

Outputs

Output files:

• gsea_pvals.csv : p-value for the enrichment test
• gsea_scores.csv : enrichment scores
• gsea_norm.csv : normalized enrichment scores