Many-Body Expansion of Direct and Mutual Polarization Models

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Teresa Head-Gordon of the University of California, Berkeley is supported by an award from the Chemical Theory, Models and Computational Methods program in the Chemistry Division to develop computational and theoretical models and methods that can be used to make accurate predictions for the behavior of complex molecular systems. They are developing and testing the accuracy of several cost-effective approximate models. The goal is to understand when these approximate models are adequate and when more costly and accurate models are required. Professor Head-Gordon and her coworkers use their methods to study the properties and behavior of several important molecular and biomolecular systems including supercooled water, ionic liquids and proteins. Ionic liquids are more environmentally benign than traditional solvents used in industrial processes. For example, ionic liquids are able to dissolve biological plant materials so that they may be converted into renewable energy products. The Head-Gordon group also studies biological systems ranging from small peptides in water to the formation of protein-protein and protein-nucleic acid interactions. Such interactions are required for diverse biological processes such as signaling, regulation of gene expression, cell division, and DNA repair. An additional goal of this project is to foster opportunities for undergraduate research and to mentor women in the physical sciences. The Head-Gordon research group has developed the Many-Body Expansion (MBE) of polarization to tackle the characterization of large condensed phase systems with a more accurate force field. In principle, mutually polarizable models offer a significant improvement in the physics of classical force fields. However, the corresponding increased computational cost renders statistical convergence of condensed phase observables more difficult to achieve. The researchers are analyzing the quality of the approximation to the MBE truncated at 3-bodies for four different polarizable models: iAMOEBA: an atomistic model that only accounts for direct polarization; AMOEBA: an atomistic point dipole model that accounts for mutual polarization; PB-SAM: a new semi-analytical solution to the linearized Poisson-Boltzmann equation (PBE) that accounts for complete mutual polarization; and iPB-SAM: devised to account for direct polarization only. The new atomistic-based polarizable models and algorithms provide a more sophisticated level of model physics for bulk water and ionic liquids as well as peptides and proteins. New PB-SAM formulations of the PBE allow simulation of the mechanism of complex formation and rate of association needed for understanding function in the crowded cellular environment, for example ~1000 proteins in E. coli cytoplasm or other complex materials such as Nafion. In addition it provides a practical guide to consumers of chemical software in the proper application of theoretical models, as to which statistical properties and mechanisms of condensed phase chemical systems actually require an advancement over a classical pairwise additive force field.