Supplementary files for "Learning the Relationship Between Variants, Metabolic Fluxes and Phenotypes"

Our study presents a novel approach, VaMP, which integrates neural networks and metabolic models to understand the complex links between genetic variants and cancer phenotypes. By leveraging genome-scale metabolic models (GSMs) and neural networks, VaMP offers a comprehensive framework for decoding cancer biology. We demonstrate its effectiveness through validation experiments on five cancer types, showcasing its ability to identify significant genes and metabolic signatures.