Modeling Molecular Recognition with the Polarizable Gaussian Multipole Model

Atomistic molecular simulations offer invaluable insights into the mechanisms underlying various diseases, enabling the generation of testable hypotheses. However, accurately modeling complex processes such as order-disorder transitions, ionic interactions, and interactions at the biomembrane interface remains challenging due to the need for precise representation of electrostatic and polarization effects across diverse structural states and solvent environments.

To address this, we are developing the polarizable Gaussian Multipole (pGM) model. By representing charges and multipoles using Hermite-Gaussian functions, the pGM model improves accuracy, self-consistency, and transferability compared to traditional delta function-based approaches. Our recent efforts have focused on designing and implementing the functional form and parameterization schemes for the pGM model, optimizing isotropic atomic polarizabilities and radii, and establishing a local reference frame that enables closed-form analytical expressions for atomic forces. For molecular simulations under periodic boundary conditions, we have interfaced the pGM model's electrostatic terms with the Particle Mesh Ewald approach, and derived the pGM internal stress tensor expression for constant pressure MD simulations. Our analysis demonstrates the accuracy of the pGM model in predicting many-body interactions in peptide oligomers and its transferability in predicting electrostatic potentials of molecules and oligomers.

Despite their potential, polarizable models have been limited by computational inefficiencies. To overcome this, we are incorporating multi-scaling into the development of polarizable force fields. Specifically, we are actively exploring the integration of continuum solvent models with the pGM electrostatics framework, and developing automated and consistent coarse-grained mapping strategies with pGM-compatible potentials. These advancements in the pGM model contribute to the development of more accurate and efficient molecular simulations, ultimately facilitating a deeper understanding and more effective treatment of diseases.