A Framework Based on Metabolic Networks and Biomedical Images Data to Discriminate Glioma Grades

Collecting and integrating information from different data sources is a successful approach to investigate complex biological phenomena and to address tasks such as disease subtyping, biomarker prediction, target, and mechanisms identification. Here, we describe an integrative framework, based on the combination of transcriptomics data, metabolic networks, and magnetic resonance images, to classify different grades of glioma, one of the most common types of primary brain tumors arising from glial cells. The framework is composed of three main blocks for feature sorting, choosing the best number of sorted features, and classification model building. We investigate different methods for each of the blocks, highlighting those that lead to the best results. Our approach demonstrates how the integration of molecular and imaging data achieves better classification performance than using the individual data-sets, also comparing results with state-of-the-art competitors. The proposed framework can be considered as a starting point for a clinically relevant grading system, and the related software made available lays the foundations for future comparisons.