(403d) Methodologies for Enhanced Unbiased Sampling of the Free Energy Landscapes of Proteins

Protein function is inextricably linked to protein dynamics. As we move from a static structural picture to a dynamic ensemble view of protein structure and function, novel computational paradigms are required for observing and understanding conformational dynamics of proteins and its functional implications. In principle, molecular dynamics simulations can provide the time evolution of atomistic models of proteins, but the long time scales associated with functional dynamics make it difficult to observe rare dynamical transitions on the free energy landscapes associated with protein function. One approach to surmount this challenge is to use reaction coordinate that characterize the rare conformational transitions to sample conformational landscapes more effectively. However, it is not possible to identify reaction coordinate for systems a priori, especially for proteins with limited availability of structural information. Furthermore, different reaction coordinates characterize the free energy barriers at different points on the free energy landscapes. Here, we discuss a set of methodologies outlined in a series of papers from our group, which use evolutionarily coupling between protein residues and machine learning methods to adaptively guide sampling with no alterations to potential energy functions in order to expedite exploration of the free energy landscapes.