MSMuSig2

MSMuSig2 is a tool to identify driver genes in Microsatellite-Unstable (MSI) cancers. It analyzes recurrent MS-indel mutation counts using statistical models (Exponential, Gaussian, Weibull, Log-Normal) to find significant mutations.

---

💾 Installation

1. Clone the repo:

   git clone [https://github.com/yourusername/MSMuSig2.git](https://github.com/yourusername/MSMuSig2.git)
   cd MSMuSig2
   

2. Install R packages: Run this command in your R session:

   install.packages(c("magrittr", "minpack.lm", "ggplot2", "gridExtra", "MASS",
                      "dplyr", "data.table", "fitdistrplus", "survival", "car",
                      "ggrepel", "ggpubr", "optparse", "metap"))

---

🚀 Usage

The main script is scripts/analysis.R.

1. Input File

Your input CSV file must contain the following columns:

• CHR
• START
• PATTERN
• REFERENCE_REPEATS
• hgnc_symbol
• COUNTS

2. Run from Command Line

Rscript scripts/analysis.R --input data/your_data.csv --output results/

3. Command-Line Arguments

• -i or --input: (Required) Path to your input CSV file.
• -o or --output: (Optional) Path to the output directory. (Default: results)
• -m or --models: (Optional) Comma-separated list of models (e.g., lognormal,weibull). (Default: lognormal,exponential)
• -t or --threshold: (Optional) Minimum motif occurrence count to analyze. (Default: 45)

---

📊 Understanding the Output

All results are saved to the output directory (default: results/).

Key Outputs:

• Raw Results: <model>.csv (e.g., lognormal.csv)
  • Contains all p-values for every mutation.
• Aggregated Results: <model>_aggregated.csv
  • Contains gene-level combined p-values (using Fisher's method).
• Significant Hits: Significant_<model>.csv and Significant_<model>_aggregated.csv
  • Filtered lists of significant mutations and genes (adjusted p-value < 0.1).
• Plots: QQplot_<model>.svg and Indels_Models_fit.svg
  • Visualizations of p-value distributions and model fits.
• Model Comparison: Indels_Models_fit.csv
  • AIC/BIC scores for each model fit.

---

📄 License

This project is licensed under the MIT License.

🧑‍💻 Authors

Hagay Ladany, Dr. Yosef Maruvka.

📧 Contact

For questions, please open an issue or contact [email protected].