Charles L. Brooks III

Warner-Lambert/Parke-Davis Chair in Chemistry, Cyrus Levinthal Distinguished University

Professor of Chemistry and Biophysics, Departments of Chemistry and Biophysics, University of Michigan

Research

Novel high-throughput methods for free energy calculations

Abstract

The rapid exploration of functionally relevant chemical spaces associated with the design and refinement of small molecules as potential therapeutic agents and in the design of protein sequences in functional biologics are key to the identification of new or improved treatments for disease. Multi-site λ-Dynamics (MSλD) is a rigorous statistical mechanically framework of computationally based free energy methods that enable such calculations. MSλD facilitates the exploration of combinatorically complex chemical spaces at a level of precision that mirrors conventional pairwise free energy approaches but utilizes only a fraction of the computer wall time. In this talk, we will illustrate the basic formalism and computational infrastructure of MSλD through examples to protein-ligand optimization problems, in which both protein sequence and ligand substituents are being explored simultaneously as well as new applications to protein sensor design involving optimization of protein-protein interface interactions. These case studies will illustrate the ability to explore vast chemical spaces with the rigorous free energy methods of MSλD. Beyond the current developments within the MSλD methodology, we will describe the ability to integrate modern machine learning interfaces that present high-level quantum mechanical within the ML framework into purely Python workflows using pyCHARMM.