Ovarian Cancer Paired-end Sequencing Insert Size Analysis

This repository contains a Snakemake pipeline and accompanying scripts for analyzing shallow whole-genome sequencing (sWGS) data from ctDNA samples to differentiate between benign and malignant ovarian cancer cases using fragment length distribution.

Overview

The pipeline performs the following steps:

1. Calculate insert size metrics from BAM files for each sample.
2. Generate fragment length distribution histograms for each sample.
3. Generate a combined histogram comparing mean insert size distributions, including standard deviation, for benign and malignant cases.
4. Calculate summary statistics (mean, median, mode, and standard deviation) for each case.

Requirements

• Python 3.9
• Snakemake
• Samtools
• Conda (optional, for managing environments)

Usage

1. Clone this repository to your local machine:

git clone https://github.com/DucoG/cancer_insert_size_analysis.git

2. Modify the config_file.yaml to include the paths to your data files (BAM and BAI files) and the Conda environment file, if using Conda.

3. Install the Conda environment:

conda env create -f env.yaml

4. Activate the Conda environment:

conda activate wgs

5. Run the Snakemake pipeline with the --use-conda argument:

snakemake --cores N --use-conda

Replace N with the number of cores you want to use for the analysis.

Output

The pipeline generates the following output files:

• Insert size metrics for each sample: insert_size_metrics/
• Fragment length distribution histograms for each sample: insert_size_histograms/
• A combined histogram comparing mean insert size distributions for benign and malignant cases: combined_histogram/combined_histogram.pdf
• Plotting data (mean and standard deviation) for the combined histogram: combined_histogram/plotting_data.csv
• Summary statistics (mean, median, mode, and standard deviation) for each case: summary_statistics/summary_statistics.csv

License

MIT License