Genetic_Forecasting_of_Huntington_Disease

📖 Overview

Huntington’s Disease (HD) is a hereditary neurodegenerative disorder caused by abnormal expansions of CAG trinucleotide repeats in the HTT gene.
This project proposes a Deep Neural Network (DNN) based forecasting system that predicts the likelihood of HD using genetic sequence data.
The model analyzes DNA sequences, identifies abnormal repeat patterns, and classifies individuals as either healthy or at risk.

---

🎯 Objectives

• To build a computational model capable of detecting abnormal CAG expansions in the HTT gene.
• To leverage Deep Learning for early risk prediction of Huntington’s Disease.
• To evaluate the model using accuracy, precision, and AUC.
• To provide a framework for genomic-based clinical decision support.

---

🧪 Methodology

1. Data Collection

• DNA sequences focusing on CAG repeat regions of the HTT gene.
• Public sources: NCBI GenBank, Ensembl Genome Browser, GEO datasets.
• Example dataset (sample_data.csv) is provided with synthetic sequences.

2. Preprocessing

• One-hot encoding of DNA nucleotides (A, T, G, C).
• Normalization of sequence lengths (padding/truncation).
• Splitting into training (80%) and testing (20%) datasets.

3. Model Architecture

The DNN is based on a 1D Convolutional Neural Network (CNN):

• Conv1D – learns sequence motifs.
• MaxPooling1D – reduces noise.
• Dropout – prevents overfitting.
• Dense layers – fully connected classification.
• Sigmoid output – binary prediction (0: healthy, 1: at-risk).

4. Training

• Loss: Binary Crossentropy
• Optimizer: Adam
• Metrics: Accuracy, Precision, AUC

5. Evaluation

• Performance is measured on unseen test data.
• Metrics include:
  • Accuracy: Overall correctness.
  • Precision: Correct positive predictions (HD risk).
  • AUC: Distinguishing power between classes.

---

⭐ Give a Star!

If you like this project, feel free to ⭐ the repo.
Your support motivates me to build more projects!