Christopher Jarzynsky of the University of Maryland, College Park, is supported by an award from the Theoretical and Computational Chemistry program within the Division of Chemistry for the development of theoretical tools to improve computational efficiency in biomolecular computations. His research focuses on increasing the efficiency of two important and challenging computational problems: free energy estimation and the sampling of rugged potential energy landscapes. The methods of computational thermodynamics are based on statistical-mechanical identities which relate thermodynamic properties of systems to their underlying microscopic details. While fundamentally sound, these methods are typically computationally expensive. To enhance computational efficiency, the PI has developed a novel toolkit of strategies organized around the principle of enhancing overlap between thermodynamic states.
In addition to having a broad impact on biomolecular applications, the methods being developed are relevant in a wide range of other contexts in materials, physics, statistics, and chemistry. The PI has launched a community-driven website devoted to methods for free energy calculations.
Computer simulations have become an indispensable tool for studying complex molecular systems, such as the biological molecules that are the basis of life. The research funded by this project has focused on the development of accurate and efficient methods for estimating thermodynamic properties of such complex systems. In particular, in this research we have formulated a toolkit of strategies to improve computational efficiency, organized around the principle of enhancing overlap between thermodynamic states. This toolkit includes: (1) A novel method for using nonequilibrium simulations to improve the effectiveness of the widely used Replica Exchange Method for computational sampling. (2) A strategy for applying the Targeted Free Energy method to estimating the thermodynamic cost of growing a soluted in solvent. (3) A new approach for computing the absoluted free energy of a crystalline solid. In addition, this research has uncovered a fundamental relationship between physical and information-theoretic measures of dissipation, which contributes to our basic knowledge of systems away from thermal equilibrium. This research has contributed directly to the training of graduate students, who have sharpened their analytical and computational skills, and broadened their scientific knowledge, while tackling an intellectually challenging problem. Finally, the PI has engaged in broader activities that promote the progress of science. He has developed a one-week introductory course on nonequilibrium physics, and has presented this course at summer schools in the United States, Belgium, Spain and India. He has also co-edited a book, "Nonequilibrium Statistical Physics of Small Systems" (Wiley-VCH, 2013). Finally, he has given numerous department-wide colloquiua aimed at introducing a broad range of chemists and physicists to the exciting recent progress in fundamental nonequilibrium statistical mechanics.
