Bulk RNA sequencing pipeline

Snakemake pipeline processes paired-end FASTQ files through quality control, alignment, transcript quantification, differential expression analysis (DESeq2), and GO enrichment analysis

Pipeline Features

• FastQ Quality Control (FastQC, MultiQC)
• Adapter Trimming (Atria)
• Read Alignment (STAR)
• Read Alignment QC (MultiQC)
• Transcript Quantification (Salmon)
• Quality Control Metrics (RSeQC, samtools)
• Differential Expression Analysis (DESeq2)
• GO Enrichment Analysis (clusterProfiler)

📥 Installation

Clone the Repository

git clone https://github.com/your_username/Bulk_RNA_seq.git
cd Bulk_RNA_seq

Install Dependencies

Snakemake

conda install -c bioconda pandas snakemake=8.25.0

It is best to stick with version 8.25.0 of Snakemake as the pipeline was optimised with this version.

atria

We use atria version 4.1.1 for trimming the adapters.

Use singularity (default)

Note

Singularity/Apptainer must be install on all machines running the pipeline

Build the image

cd atria
sudo singularity build atria-4.1.1.sif atria.def
export ATRIA_IMG_DIR=$PWD

Set SINGULARITY_CACHEDIR or use --apptainer-prefix $ATRIA_IMG_DIR with the Snakemake command

export SINGULARITY_CACHEDIR=$ATRIA_IMG_DIR

Install atria locally

You install it on your machine using the script install_atria.sh in workflow/scripts.

You will need to add it to your path and your .bashrc:

export PATH=$PATH:<path to atria>/atria-4.1.1/bin

User Input Files

Sample Sheet

Currently, you can perform analysis on paired-end reads. You must provide a CSV file describing your samples. If your data are in Excel format, it can be easily save as a CSV file. The sample sheet should contain the following columns:

• group: Experimental group (e.g., cell line or condition), only used with the comparison file
• sample_id: A unique identifier for each sample, needs to match with samplename
• replicate: Replicate number (e.g., 1, 2, …), optional, this is for information only and is not used by the pipeline
• fastq_1: The absolute path to the fastq files with the first read pairs
• fastq_2: The absolute path to the fastq files with the second read pairs

Example:

group,sample_id,replicate,fastq_1,fastq_2
GM12878,SRR3192657,1,/data/myexperiment_1/SRR3192657_1.fastq.gz,/data/myexperiment_1/SRR3192657_2.fastq.gz
GM12878,SRR3192658,2,/data/myexperiment_1/SRR3192658_1.fastq.gz,/data/myexperiment_1/SRR3192658_2.fastq.gz
K562,SRR3192408,1,/data/myexperiment_2/SRR3192408_1.fastq.gz,/data/myexperiment_2/SRR3192408_2.fastq.gz
...

The default name is sample_sheet.csv and its default location is in the data folder. However, you can edit the configuration file config/config.yml

sample_sheet: "data/sample_sheet.csv"

Or you can specify the path to the file on the command line:

snakemake --software-deployment-method conda --config sample_sheet=<path/to/config>

Raw fastq files

The pipeline will use the path found in the sample sheet to locate the FASTQ files.

Comparisons File

[!INFO] If you do not want to compare the gene expression levels between samples, set do_comparisons to false but leave the default value of data/comparison.cvs in your config.

Provide a CSV file defining the comparisons for differential expression analysis. The comparisons file should include the following columns:

• comparison_number: A unique identifier for the comparison
• treatment: Name of the treatment group
• control: Name of the control group

comparison_number,treatment,control
1,GM12878,K562
2,MCF7,H1

The default name is comparison.csv and its default location is in the data folder. However, you can edit the configuration file config/config.yml

comparison_sheet: "data/comparison.csv"

Or you can specify the path to the file on the command line:

snakemake --software-deployment-method conda --config comparison_sheet=<path/to/config>

🚀 Running the Pipeline

1️⃣ Test Pipeline Execution (Dry Run)

snakemake --software-deployment-method conda -np

2️⃣ Run the Full Pipeline

Using the default Snakemake way

By default, Snakemake expects you to run the pipeline in the pipeline repository and all files and folders will be created in the pipeline directory

snakemake --software-deployment-method conda singularity --singularity-args "-B $HOME" --rerun-triggers mtime --cores 30

💡 Adjust --cores based on your system's available CPUs. It will be the limit

[!INFO] Snakemake mount the current directory in Singularity with --home. You may need to use --singularity-args to mount the directory where your input and output FASTQ files from trimming would be

Using a specific output directory

Tip

It is worth using absolute path in the config file to avoid troubles later

snakemake --software-deployment-method conda singularity --singularity-args "-B $HOME" --rerun-triggers mtime --directory <path to your output files> --configfile $PWD/config/config.yaml

The following parameters would need to be update in the config or with --config KEY=VALUE

• sample_sheet
• comparison_sheet
• atria_image

Pipeline options

• trim_reads: Set to false if reads do not need to be trimmed. The trimming software, atria, is doing other cleaning, it might be worth keeping it. If it doesn't find adapters, it doesn't remove them.
• do_comparisons: Set to false if you do not want to run DESeq2 and GO enrichments analyses

Useful snakemake parameters

• --directory PATH: set the pipeline directory to the path provided
• --configfile PATH: set the path to the config file to use, needed if using --directory or you will have problems with relative paths
• --snakefile PATH: set the path to the Snakemake file to use, needed if you want to run the pipeline outside of the pipeline repository.
• --config KEY=VALUE: set a value from the config file with the CLI
• --max-threads: Limit the number of threads to use