Molecular Analysis based on Genetic Feature Selection for Gliomas Classification

Genetic testing is a rapidly evolving modality in cancer management, with DNA microarrays facilitating computational analyses at the molecular level. However, current computational methods often fail to account for gene interactions, limiting their ability to elucidate the underlying pathogenesis. This study introduces a heuristic feature selection algorithm that effectively categorizes glioma grades by identifying specific gene subsets. Our discretization technique addresses the challenge of large-scale microarray data, and by focusing on gene combinations rather than individual features, the algorithm provides insights into the critical roles of these subsets in glioma pathogenesis.