This repo contains several workflows to perform several tasks:

1. fastq2counts, which is our configuration of nf-core/rnaseq to generate count matrices from fastq files
2. de_analysis to perform differential expression analysis with deseq2 using negative binomial distribution or limma with linear model using the count matrices

fastq2counts

nf-core/configs: UzL OMICS Cluster Configuration

The rnaseq nf-core pipeline has been successfully configured for use on the UzL OMICS cluster at the University of Luebeck.

To use, run any nf-core pipeline with -profile uzl_omics. This will download and launch a copy of the UzL OMICS cluster specfic configuration which has been uploaded to nf-cores repository of many configurations for institutes around the world. Using this profile, docker images containing the required softwares will be downloaded, and converted to singularity images before execution of the pipeline.

Software dependencies

Nextflow and Singularity are required to run nf-core scripts. Singularity is available using the environment module system on UzL OMICS cluster and can be made available by issuing the command below:

module load singularity

But nf-core requires Nextflow version 22.10.1 or higher, so you have to install a more recent version first (as of Nov 2023, higher versions are not installed in the cluster). This can be done within a conda environment for example. Note: for Nextflow versions newer than 23.07.0-edge, it is necessary to mount the home directory using the command:

NXF_SINGULARITY_HOME_MOUNT=true

Running the pipeline using this configuration

You can use the following batch script as an example to run the pipeline (the example shown for the nf-core/rnaseq pipeline can also be extended to other nf-core pipelines):

#!/bin/bash

#SBATCH -c 1
#SBATCH --mem=10GB
#SBATCH --partition=shortterm

# Make conda available within a slurm job within which Nextflow is installed
PATH=$WORK/.omics/miniforge3/bin:$PATH
source $WORK/.omics/miniforge3/etc/profile.d/conda.sh
conda activate YOUR_NEXTFLOW_ENV     #you have to activate your environment with a Nextflow version 22.10.1 or higher


module load singularity

cd PATH_TO_YOUR_LAUNCHDIR
export NXF_SINGULARITY_HOME_MOUNT=true
nextflow run nf-core/rnaseq \
    -profile uzl_omics \
    -params-file PATH_TO_YOUR_PARAMS_FILE

Below are non-mandatory information

note: You will need access to the UzL OMICS cluster in order to run the pipeline. In doubt contact IT.

de_analysis

Two options are available for the analysis of diffferentially expressed genes, DESeq2 and Limma. These can be run as described below.

deseq2

Fill out the parameters file start the job as follows:

Rscript deseq2.R

By default the R script deseq2.R uses the parameters from params_deseq2.yaml to execute deseq.rmd.

This gives an output folder in the provided outdir named after the comparison provided in the parameters file. As output you should receive the DESeq2 report as an an HTML file it this ordner, as well as the used parameters. If wanted, .xlsx files including the DESeq2 results for all genes and especially DE genes can be generated before and after potential batch correction.

note: Make sure to use an unnormalized counts table.

limma

Similar to the execution of deseq2, the parameters file must be customized. Start the job with the following command:

Rscript limma.R

By default the R script limma.R uses the parameters from params_limma.yaml to execute limma.rmd.