Protein conformational dynamics are essential for biological function. While AlphaFold can predict native protein structures, it cannot capture alternative conformations or conformational changes. We developed a rigorous, physics-based framework—the generalized work functional—that enables to directly compute, from energy relaxation simulations, the true reaction coordinates: the few essential protein coordinates that fully determine protein conformational dynamics and accurately predict the committor for any given conformation. The reaction coordinates enable efficient and predictive simulations of large-scale, functionally important conformational transitions and the resulting alternative conformations. The method’s accuracy and generality are demonstrated across diverse protein systems, including HIV-1 protease and reverse transcriptase, the human PDZ domain, and major histocompatibility complex I.
Ao Ma is an Associate Professor of Biomedical Engineering at the University of Illinois Chicago. His research combines computational modeling, statistical physics, and molecular biophysics to study how proteins and cellular structures work and how their functions change in disease. His group develops theoretical and computational tools to uncover the molecular mechanisms of large-scale protein conformational dynamics and cytoskeletal dynamics. He is also active in building interdisciplinary research communities and mentoring young scientists in quantitative biology.