DNA Classifier using Deep Learning

This project builds a deep learning model to classify synthetic DNA sequences as either human or non-human using k-mer frequency encoding.

Project Structure

notebook/ → Jupyter Notebook with full pipeline data/ → Generated synthetic dataset (CSV) models/ → Trained Keras model (.keras) + scaler (Pickle)

Model Summary

• Data: 1000 synthetic DNA sequences (500 human, 500 non-human)
• Encoding: 3-mer frequency vector (64 features)
• Model: Feedforward Neural Network using TensorFlow/Keras
• Accuracy: 100% test accuracy on clearly separated synthetic data

How to Run

# Create virtual environment
conda create -n dnaenv python=3.10
conda activate dnaenv

# Install libraries
pip install -r requirements.txt

# Run the notebook
jupyter notebook notebook/dna_classifier_final.ipynb