Ensemble Machine Learning Approaches Predict Survival in Lower-Grade Glioma Based on Glycosphingolipid Gene...

Glycosphingolipids (GSLs) are essential components of biological membranes with important roles in cell signaling. Disrupted GSL metabolism is associated with malignancy across a range of cancers, with different GSLs implicated in distinct tumors. GSLs have potential mechanistic roles in cancer; however, their functions in lower-grade gliomas (LGGs) remain poorly understood. We present ensemble machine learning approaches using transcriptomic data from LGG, combined with GSL-specific metabolic simulations, to predict survival outcomes. The ensemble approach demonstrates effective risk stratification for LGG patients based on GSL synthetic enzyme expression. Pathway analysis of model-derived risk groups identified correlations with GSL-modulated pathways including cell motility, division, and Wnt signaling in LGG pathology. Given the strong performance of machine learning approaches to predict survival outcomes and that GSLs are shed into the tumor microenvironment, GSL-based diagnostics and prognostics may prove to be clinically beneficial upon future experimental validation. A Python package enabling GSL-specific metabolic modeling and risk prediction from RNA sequencing data is provided.