Description

This pipeline performs differential gene expression and gene set enrichment in droplet single cell RNA-seq. As input, this pipeline takes an H5AD single cell experiment, modeled after the output of the complimentary pipeline nf_scrna_qc.

For differential gene expression, we support the following packages:

1. MAST
2. edgeR
3. DESeq2

For gene set enrichment, we support the following packages:

1. fGSEA

For multiple correction, we support:

1. Benjamini-Hochberg
2. IHW

For a full description of possible parameters, see README-params.md.

For more information on running the pipeline, see:

• Common problems
• Output directory structure

Quickstart

Quickstart for deploying this pipeline locally and on a high performance compute cluster.

1. Set up the environment

See environment README to set up environment. Once the environment is set up, activate the conda environment:

source activate sc_diff_expr

Alternatively, if using singularity or docker, one can pull the image from letaylor/sc_nf_diffexpression:latest.

2. Prepare the input files

Generate and/or edit input files for the pipeline.

The pipeline takes as input:

1. --file_anndata: H5AD file containing sequencing data. Required.
2. -params-file: YAML file containing analysis parameters. Required.

Examples of these files can be found in demo/.

3. Run pipeline

NOTE: All input file paths should be full paths.

To run:

nextflow run \
  "/path/to/repo/dir/main.nf" \
  -profile "local" \ # could also be configuration for cluster
  --file_anndata "/path/to/data/data.h5ad" \
  -params-file "/path/to/config/params.yml"

Examples:

• Local with docker
• Local with singularity
• Cluster with conda

Authors: Henry Taylor and Leland Taylor