ML-Prediction-of-Diabetes-Biomarker-

This project investigated the potential of machine learning in complementing established techniques for identifying key genes involved in Type 2 Diabetes, a condition with a complex genetic background. DGE analysis showed significant dysregulation in 746 out of 18,484 genes. Machine learning models, specifically SVM, RF, LR, and XGBoost were evaluated on the RNA-Seq count dataset. The decision tree-based algorithm XGBoost outperformed other classifiers and SHAP analysis of this model helped identify top key genes. Raw count file: 2117ensemble.csv Clinical metadat: clin2117.csv GEO accession of dataset: GSE182117