Hybrid simulations: combining atomistic and coarse-grained force fields using virtual sites

Abstract

Hybrid simulations, in which part of the system is represented at atomic resolution and the remaining part at a reduced, coarse-grained, level offer a powerful way to combine the accuracy associated with the atomistic force fields to the sampling speed obtained with coarse-grained (CG) potentials. In this work we introduce a straightforward scheme to perform hybrid simulations, making use of virtual sites to couple the two levels of resolution. With the help of these virtual sites interactions between molecules at different levels of resolution, i.e. between CG and atomistic molecules, are treated the same way as the pure CG-CG interactions. To test our method, we combine the Gromos atomistic force field with a number of coarse-grained potentials, obtained through several approaches that are designed to obtain CG potentials based on an existing atomistic model, namely iterative Boltzmann inversion, force matching, and a potential of mean force subtraction procedure (SB). We also explore the use of the MARTINI force field for the CG potential. A simple system, consisting of atomistic butane molecules dissolved in CG butane, is used to study the performance of our hybrid scheme. Based on the potentials of mean force for atomistic butane in CG solvent, and the properties of 1 : 1 mixtures of atomistic and CG butane which should exhibit ideal mixing behavior, we conclude that the MARTINI and SB potentials are particularly suited to be combined with the atomistic force field. The MARTINI potential is subsequently used to perform hybrid simulations of atomistic dialanine peptides in both CG butane and water. Compared to a fully atomistic description of the system, the hybrid description gives similar results provided that the dielectric screening of water is accounted for. Within the field of biomolecules, our method appears ideally suited to study e.g. protein-ligand binding, where the active site and ligand are modeled in atomistic detail and the rest of the protein, together with the solvent, is coarse-grained.

Computer Simulations of Soft Matter: Linking the Scales

“…Recently Noid and coworkers have analyzed these aspects using concepts from liquid state theory [100,120]. One more note concerning Henderson's theorem: even though there is in principle one exact solution for the effective pair potential that reproduces a given pair correlation function, different potentials might give a reasonably close representation of the structure, i.e., the above inverse problem is mathematically ill-posed [116,121].This effect becomes even more pronounced in complex systems where several interaction functions corresponding to several RDFs need to be numerically determined. This can to some extent be turned into an advantage since it allows one to impose thermodynamic constraints in the parametrization procedure.…”

“…In addition, it is of particular methodological interest because for single bead models of water it is known that three-body correlations play a decisive role and the potential compromise between reproducing pair-or higher order structural correlations is particularly relevant for the properties of the model [101,116,117]. Different studies have been carried out that compare structure-motivated and thermodynamics-based CG models [121,136,137]. While CG models where the parametrization targets had been solvation and partitioning properties are particularly well suited to reproduce processes where for example hydrophilicity/hydrophobicity arguments play a decisive role, they do not per se reproduce the structure of the system [121,136].…”

“…Different studies have been carried out that compare structure-motivated and thermodynamics-based CG models [121,136,137]. While CG models where the parametrization targets had been solvation and partitioning properties are particularly well suited to reproduce processes where for example hydrophilicity/hydrophobicity arguments play a decisive role, they do not per se reproduce the structure of the system [121,136]. Related to their ability to reproduce the thermodynamic properties of certain chemical units, these models exhibit considerable transferability and can often be applied to a variety of molecular systems and a range of thermodynamic conditions.…”

Computational ‘microscopy’ of cellular membranes

“…Hybrid models are, in principle, a very powerful simulation strategy because they combine the best of both worlds -sampling globally at a low resolution (for instance, the bulk membrane and solvent) with only a local high resolution in defined areas of interest (for instance, at the protein-lipid interface). Either a static division of the all-atom and coarse-grained degrees of freedom can be used (Han and Schulten, 2012;Rzepiela et al, 2011;Wassenaar et al, 2013), akin to the wellestablished hybrid quantum mechanics/molecular mechanics (QM/ MM) method, or an adaptive boundary that allows particles to change resolution during the simulation (Praprotnik et al, 2008;Zavadlav et al, 2014). A number of pioneering hybrid studies of the static kind have appeared, in which all-atom/based membrane proteins are embedded in a coarse-grained model membrane environment (Han and Schulten, 2012;Wassenaar et al, 2013).…”

Adaptive resolution simulation of an atomistic DNA molecule in MARTINI salt solution

“…However, it typically comprises most of the particles in the simulated system and thus the majority of the computational effort is spent on obtaining water-water interactions in distal regions that are not relevant for the problem under consideration. This realization led to the development of several multiscale simulation methods, which reduce the number of degrees of freedom for distal water and at the same time keep the atomistic (AT) resolution where it is necessary [5][6][7][8][9]. The speedup of such multiscale simulations with respect to all-atom simulations is proportional to the reduction of the interaction sites in the coarse-grained (CG) model of water.…”