ANALYSIS FOR DISEASE GENE ASSOCIATION USING MACHINE LEARNING

The project "Analysis for Disease Gene Association Using Machine Learning" aims to analyze and predict associations between genes and diseases using advanced machine learning techniques. With the growing availability of genetic data, understanding how specific genes relate to diseases has become vital in biomedical research. This study utilizes a comprehensive dataset that includes gene-specific information such as Disease Specificity Index (DSI) and Disease Pleiotropy Index (DPI), alongside various disease features like semantic type and classification.

The project implements four machine learning algorithms—Random Forest, XGBoost, LightGBM, and K-Nearest Neighbors (KNN)—to predict three key output classes: Disease, Group, and Phenotype. The Random Forest model achieved the highest accuracy (97.81%), and it was deployed using a Flask framework for real-time predictions. Preprocessing steps included filling missing values, label encoding, and clustering to categorize diseases. KMeans clustering grouped diseases into broader categories based on similarities, further enhancing the prediction capabilities.

The project demonstrates the potential of machine learning in advancing genomic research by providing insights into gene-disease associations. It offers a practical tool for researchers to explore genetic links to diseases efficiently.

Keywords: gene-disease association, machine learning, clustering, Flask, Random Forest, disease classification, phenotype prediction, gene prediction, bioinformatics, data preprocessing