Will predictive AI make discovery research in metabolism obsolete

Human metabolism operates as a deeply interconnected network that maintains metabolic homoeostasis by coordinating glucose and lipid utilization, amino acid synthesis, and other core biochemical processes with the handling of exogenous compounds such as drugs and environmental xenobiotics. These endogenous and exogenous pathways are tightly integrated: fluctuations in nutrient availability, alterations in lipid or glucose flux, and exposure to chemicals or therapeutic agents can all reprogramme metabolic circuits, shifting enzyme activities, transporter function, and regulatory signalling 1 . Such cross-talk between endogenous and exogenous metabolic pathways not only influences systemic energy balance and cardiometabolic health, but also modifies the absorption, distribution, metabolism, and excretion of xenobiotics. As a result, understanding how metabolic homoeostasis governs—and is perturbed by—this bidirectional interplay has become essential for dissecting disease mechanisms, anticipating adverse metabolic responses, and improving the safety and efficacy of drug development pipelines.