Co-mutation Scripts

Introduction

This project provides a set of scripts for mining of co-occurring mutation pairs, which includes steps for data preprocessing, applying the Frequent Pattern Growth (FP-Growth) algorithm, calculating co-scores, and computing the Jaccard index and Fisher's exact test.

File Description

0.SSM_preprocessing

• 1.ssm_liftover_and_filter_repeat.sh: - Use liftOver to convert the genome version; Filter simple somatic mutations (SSMs) located in repeat regions
• 2.ssm_data_arrange.py: Filter out cancer types with fewer than 100 donors; Split the SSMs data by cancer type to prepare the input for the FP-growth algorithm.

1.FP-growth

• fp_cancers.py: FP-growth input arrangement and running.
• fp_cancers.sh: Template to use the fp_cancers.py
• batch_submit_cancertype.sh: Submit FP-growth tasks in batches by cancers using the slurm job system.

2.Statistics of second-order mutation pairs

• 1.fisher_J_mut2.py: Calculate Jaccard index and Fisher's exact test for second-order mutation pairs.
• 2.fisher_J_mut2_batch.sh: Batch submit 1.fisher_J_mut2.py script for each cancer.
• 3.co-scores_calcu.py: Calculate co-scores for each second-order mutation pair.
• 4.co-scores_calcu_batch.sh: Batch submit 3.co-scores_calcu.py script for each cancer.

3.Statistics of third-order mutation pairs

• 1.mut3_supp_score.py: Calculate co-scores for each third-order mutation pair.
• 2.mut3_supp_score.sh: Batch submit 1.mut3_supp_score.py script for each cancer.

Dependencies

• Python 3.x
• Required Python modules: pandas, numpy, subprocess, argparse, os, scipy.
• Please install FP-growth from: https://borgelt.net/doc/fpgrowth/fpgrowth.html
• Unix/Linux operating system.

Citation

Coming soon...