🐱🧬 CatCRISPR: Predicting sgRNA Efficiency with CatBoost

This project focuses on predicting sgRNA efficiency scores using CatBoost, a gradient boosting algorithm on decision trees, with one-hot encoded sequence features.

📋 Project Overview

Designing efficient sgRNAs is essential for accurate CRISPR-Cas9 genome editing.
CatCRISPR applies machine learning to predict sgRNA efficiency by:

• Encoding sequences into one-hot vectors
• Training a CatBoost regression model
• Evaluating performance with cross-validation and regression metrics

📁 Project Structure

• Code/ – Source code for preprocessing, fine tunning, and training
• Data/ – Input datasets of sgRNA sequences and labels

🚀 Key Features

1. 🐱 CatBoost regression for sgRNA efficiency prediction
2. 🔡 One-hot encoding of sgRNA sequences
3. 🔄 5-fold cross-validation for robust evaluation
4. 📉 Regression metrics: Spearman, Pearson correlation
5. 📊 Feature importance analysis to understand sequence influence

💻 Installation

Clone the repository and install dependencies:

git clone https://github.com/Sizerta/CatCrispr
cd CatCRISPR
pip install -r requirements.txt