Communication: Adaptive boundaries in multiscale simulations

Abstract

Combined-resolution simulations are an effective way to study molecular properties across a range of length and time scales. These simulations can benefit from adaptive boundaries that allow the high-resolution region to adapt (change size and/or shape) as the simulation progresses. The number of degrees of freedom required to accurately represent even a simple molecular process can vary by several orders of magnitude throughout the course of a simulation, and adaptive boundaries react to these changes to include an appropriate but not excessive amount of detail. Here, we derive the Hamiltonian and distribution function for such a molecular simulation. We also design an algorithm that can efficiently sample the boundary as a new coordinate of the system. We apply this framework to a mixed explicit/continuum simulation of a peptide in solvent. We use this example to discuss the conditions necessary for a successful implementation of adaptive boundaries that is both efficient and accurate in reproducing molecular properties. Published by AIP Publishing. https://doi.org/10.1063/1.5025826Multiscale models are an effective way to simulate molecular systems. The motivation is clear: a high-resolution model can capture physical detail, while a low-resolution model offers computational efficiency and is sometimes better suited or more easily parameterized for large-scale phenomena. 1 There are two strategies for multiscale modeling. In the first, multiple independent simulations are used for different levels of resolution. Combined intelligently, the sum is worth more than the parts: simulations at one level motivate and parameterize simulations at another. [2][3][4] The focus of this work is the second approach, which combines multiple levels of resolution into a single simulation. These simulations use a fine-grained (FG) model for a region of interest and a computationally efficient coarse-grained (CG) model elsewhere. Examples include mixed quantum and molecular mechanical, 5,6 mixed all-atom and coarse grained (CG), 7-13 and hybrid explicit-continuum solvent models. [14][15][16][17][18][19][20][21] Accurately modeling the boundary between high-and low-resolution regions is the crux of a combined-resolution simulation. Even for a homogeneous system like bulk solvent, equilibrium properties emerge from a delicate balance of interactions with the surrounding medium. An improper handling of the boundary will break the natural symmetry, and the resulting structural artifacts can extend well beyond the boundary into other regions of the simulation. 15 We have developed a hybrid explicit-continuum solvent model that includes a boundary region over which molecules gradually, rather than abruptly, change resolution. This boundary method avoids the structural artifacts common to hybrid solvent models and accurately reproduces thermodynamic properties throughout the entire explicit domain. 15,16 This boundary method is similar to that introduced in the Adaptive Resolution (AdRes) approach, 22,23 a method used to couple high r...

Microsecond Molecular Dynamics Simulations of Proteins Using a Quasi-Equilibrium Solvation Shell Model

“…Alternatively, a more elaborate solvent boundary potential could be developed, e.g., one with an outward repulsion force to counteract the surface tension phenomenon . Another possibility is to generalize the equilibrium formalism of Wagoner and Pande developed for multiscale simulations, in which the solvent boundary is treated as an explicit coordinate of the simulation system; notably, the present version of the model has only been tested on peptides solvated in spherical droplets …”

From System Modeling to System Analysis: The Impact of Resolution Level and Resolution Distribution in the Computer-Aided Investigation of Biomolecules

“…Attempts of triple-scale simulation of liquid water have also been performed, by concurrently coupling atomistic, CG, and continuum models ( Delgado-Buscalioni et al, 2009 ; Zavadlav et al, 2018 ). Applications of an atomistic/continuum representation of the solvent for biomolecular studies have been performed in Wagoner and Pande (2018) , where the boundary between the explicit/continuum solvent models can adapt itself in response to the conformational fluctuations of the atomistic peptide simulated; and in Hu et al (2019) , for a multi-resolution simulation of protein diffusion in water under a steady shear flow.…”

Atomistic Simulations of Membrane Ion Channel Conduction, Gating, and Modulation

“…One can increase the explicit representation of the bulk solvent or move to new methods that can adjust the system size on the fly. 96 The more commonly used approach, however, is to impose periodic boundary conditions (PBC), where as a particle leaves the central image it is translated and enters from the opposite face. 84 However, it remains important to simulate a large enough box of the system to avoid artifacts from the boundaries.…”