Application of Machine Learning Techniques to Drive Immunological Insights Towards Malaria Prognosis Using Microarray data

This study aims to identify genes related to malaria using a comprehensive approach that leverages multiple machine learning-based feature selection techniques. Due to the complexity and resource intensity of building weighted gene co-expression network analysis (WGCNA) networks, we utilized various machine learning algorithms to streamline the process. We analyzed the GSE117613 dataset from the Gene Expression Omnibus (GEO) to identify differentially expressed genes (DEGs) between malaria-infected and control samples. Techniques such as the Least Absolute Shrinkage and Selection Operator (LASSO), Sup-port Vector Machine Recursive Feature Elimination (SVM - RFE), SVM with Radial Basis Function (RBF), ensemble model, ReliefF, and SigFeature were employed to select potential biomarkers. By intersecting the genes captured by these methods, we identified robust candidates for further analysis. Pathway analysis was then conducted to elucidate the biological significance of these genes in the context of malaria. Our results highlighted several significant genes associated with malaria, validated using pathway analysis. This study provides insights into the molecular mechanisms of malaria and demonstrates an efficient, machine learning-based approach for biomarker discovery in medical research.