Proteins conform the ultimate machinery of Life, executing all processes that keep us alive– from metabolic pathways to neurotransmission. Behind every nutrient absorbed, or even every sensation, there is buzzling protein activity, carefully polished by millions of years of evolutionary selection. To efficiently push sugar through a membrane, transport an ion, every protein has refined its shape to perform a specific task. As sequences fold into 3D-sructures, these encode intrinsic functional motions (aka “conformational changes”) - the ultimate matter of evolution. Hence, mutations that disrupt native motions cause disease, as is well-known for oncogenes. Despite its importance, understanding the link between structure, motion and function is a challenge. It is extremely hard to trap protein structures in action, and also, to model them. Conformers captured experimentally are stable end states, but short-lived flexible intermediates—key for molecular mechanisms—are elusive due to the “sampling” problem. We argue there is a golden mine of conformational structure-function information, hidden in the genomic screenings for cancer and mendelian diseases. As of 2024, there are over 14 million missense mutations in the COSMIC cancer database. It is assumed that a majority of them are just “passengers”, but our preliminary modelling studies, challenge this. We propose than more mutations than ever imagined are not random, but precisely selected to fine-tune every possible gear in cells and moreover, can unveil unexpected conformational landscapes. Our laboratory is developing an ambitious multiscale simulation workflow to dive in this mutational sea in search of “hot” mutations, with the goal to gain new insights on protein function and human disease.