Nonbonded Potentials Research Articles

Incorporating Noncovalent Interactions in Transfer Learning Gaussian Process Regression Models for Molecular Simulations.

FFLUX is a quantum chemical topology-based multipolar force field that uses Gaussian process regression machine learning models to predict atomic energies and multipole moments on the fly for fast and accurate molecular dynamics simulations. These models have previously been trained on monomers, meaning that many-body effects, for example, intermolecular charge transfer, are missed in simulations. Moreover, dispersion and repulsion have been modeled using Lennard-Jones potentials, necessitating careful parametrization. In this work, we take an important step toward addressing these shortcomings and show that models trained on clusters, in this case, a dimer, can be used in FFLUX simulations by preparing and benchmarking a formamide dimer model. To mitigate the computational costs associated with training higher-dimensional models, we rely on the transfer of hyperparameters from a smaller source model to a larger target model, enabling an order of magnitude faster training than with a direct learning approach. The dimer model allows for simulations that account for two-body effects, including intermolecular polarization and charge penetration, and that do not require nonbonded potentials. We show that addressing these limitations allows for simulations that are closer to quantum mechanics than previously possible with the monomeric models.

Structural Coarse-Graining via Multiobjective Optimization with Differentiable Simulation.

In the realm of multiscale molecular simulations, structure-based coarse-graining is a prominent approach for creating efficient coarse-grained (CG) representations of soft matter systems, such as polymers. This involves optimizing CG interactions by matching static correlation functions of the corresponding degrees of freedom in all-atom (AA) models. Here, we present a versatile method, namely, differentiable coarse-graining (DiffCG), which combines multiobjective optimization and differentiable simulation. The DiffCG approach is capable of constructing robust CG models by iteratively optimizing the effective potentials to simultaneously match multiple target properties. We demonstrate our approach by concurrently optimizing bonded and nonbonded potentials of a CG model of polystyrene (PS) melts. The resulting CG-PS model effectively reproduces both the structural characteristics, such as the equilibrium probability distribution of microscopic degrees of freedom and the thermodynamic pressure of the AA counterpart. More importantly, leveraging the multiobjective optimization capability, we develop a precise and efficient CG model for PS melts that is transferable across a wide range of temperatures, i.e., from 400 to 600 K. It is achieved via optimizing a pairwise potential with nonlinear temperature dependence in the CG model to simultaneously match target data from AA-MD simulations at multiple thermodynamic states. The temperature transferable CG-PS model demonstrates its ability to accurately predict the radial distribution functions and density at different temperatures, including those that are not included in the target thermodynamic states. Our work opens up a promising route for developing accurate and transferable CG models of complex soft-matter systems through multiobjective optimization with differentiable simulation.

Development of a coarse-grained molecular dynamics model for poly(dimethyl-co-diphenyl)siloxane.

Polydimethylsiloxane is an important polymeric material with a wide range of applications. However, environmental effects like low temperature can induce crystallization in this material with resulting changes in its structural and dynamic properties. The incorporation of phenyl-siloxane components, e.g., as in a poly(dimethyl-co-diphenyl)siloxane random copolymer, is known to suppress such crystallization. Molecular dynamics (MD) simulations can be a powerful tool to understand such effects in atomistic detail. Unfortunately, all-atomistic molecular dynamics (AAMD) is limited in both spatial dimensions and simulation times it can probe. To overcome such constraints and to extend to more useful length- and time-scales, we systematically develop a coarse-grained molecular dynamics (CGMD) model for the poly(dimethyl-co-diphenyl)siloxane system with bonded and non-bonded interactions determined from all-atomistic simulations by the iterative Boltzmann inversion (IBI) method. Additionally, we propose a lever rule that can be useful to generate non-bonded potentials for such systems without reference to the all-atomistic ground truth. Our model captures the structural and dynamic properties of the copolymer material with quantitative accuracy and is useful to study long-time dynamics of highly-entangled systems, sequence-dependent properties, phase behaviour, etc.

Temperature-Transferable Coarse-Grained Models for Volumetric Properties of Poly(lactic Acid).

A new coarse-grained (CG) model, for which each monomer is mapped as one bead at its center of mass, was developed for simulating the volumetric properties of the polylactide (PLA) bulk. The three bonded CG potentials are first parametrized against the strain energies of the dimer, trimer, and tetramer, and the nonbonded CG potentials are then optimized to match the melt densities of the decamer. With the derived CG potentials, molecular dynamics (MD) simulations are found to reproduce thermal expansion and glass transition. By rescaling the dihedral and nonbonded potentials with temperature-independent factors, the glass transition temperature (Tg) is also satisfactorily restored with little modifications on the volumetric expansive coefficients at both rubbery and glassy states. Therefore, the finally optimized CG potentials exhibit excellent temperature transferability, as rationalized by the Simha-Boyer relation. Furthermore, it is confirmed that the dihedral torsions and nonbonded interactions play key roles in glass transition. Also, the simulated bulk moduli and conformational properties in a wide temperature range compare well with the referenced data. The proposed multiscale scheme has great potential in simulating thermo-mechanical properties of PLA and other polymers.

Can a coarse-grained water model capture the key physical features of the hydrophobic effect?

Coarse-grained (CG) molecular dynamics can be a powerful method for probing complex processes. However, most CG force fields use pairwise nonbonded interaction potentials sets, which can limit their ability to capture complex multi-body phenomena such as the hydrophobic effect. As the hydrophobic effect primarily manifests itself due to the nonpolar solute affecting the nearby hydrogen bonding network in water, capturing such effects using a simple one CG site or “bead” water model is a challenge. In this work, we systematically test the ability of CG one site water models for capturing critical features of the solvent environment around a hydrophobe as well as the potential of mean force (PMF) of neopentane association. We study two bottom-up models: a simple pairwise (SP) force-matched water model constructed using the multiscale coarse-graining method and the Bottom-Up Many-Body Projected Water (BUMPer) model, which has implicit three-body correlations. We also test the top-down monatomic (mW) and the Machine Learned mW (ML-mW) water models. The mW models perform well in capturing structural correlations but not the energetics of the PMF. BUMPer outperforms SP in capturing structural correlations and also gives an accurate PMF in contrast to the two mW models. Our study highlights the importance of including three-body interactions in CG water models, either explicitly or implicitly, while in general highlighting the applicability of bottom-up CG water models for studying hydrophobic effects in a quantitative fashion. This assertion comes with a caveat, however, regarding the accuracy of the enthalpy–entropy decomposition of the PMF of hydrophobe association.

MolTwister – a molecular systems construction, manipulation and statistical mechanical calculation tool

To perform molecular dynamics (MD) simulations, Monte Carlo (MC) simulations, quantum mechanical (QM) electronic structure calculations, or similar atomistic calculations, it is necessary to first construct and define the molecular system of interest. This involves creating an initial configuration of atoms, where MD and MC simulations require forcefield assignments, for example in the form of non-bonded, bonded, angular and dihedral potentials. Once simulations or calculations have been performed, large sets of data are available (often several gigabytes). These contain atomic trajectories and other relevant static or dynamic information, from which static and dynamic properties (such as density profiles, vibrational density of states and velocity autocorrelation functions) can be obtained through statistical mechanical calculations. MolTwister is an open source software platform that addresses the construction of molecular systems, basic 3D visualization of these, the generation of input files for selected MD packages, as well as calculation of properties from atomistic simulation data. It also contains a GPU accelerated MD simulator suited for smaller tasks such as molecular thermalization. The software package is written in C++14 and can be used as a basis for further development, where efforts have been made to make access to underlying functionality easy. Moreover, it supports Python, where scripts have access to the majority of program functionality. Program summaryProgram Title: MolTwisterCPC Library link to program files:https://doi.org/10.17632/ph6pwst9mn.1Developer's repository link:http://moltwister.comLicensing provisions: GPLv3 or laterProgramming language: C++14Nature of problem: Lack of SW tools for easy construction of complex molecular systems, as well as tools enabling analysis of long molecular trajectories.Solution method: Developed command line tool with associated 3D view, where molecular construction commands can be executed in complex manners through a Python interface. The program also has commands for analysis of molecular data that process one molecular time-frame at a time, thus enabling analysis of long molecular trajectories. Moreover, the program structure is designed for easy implementation of new commands by other researchers.Additional comments including restrictions and unusual features: Supported file formats: xyz, pdb, mtt (MolTwister trajectory), dcd, xtc, qepos, python, script (MolTwister scripts)Supported MD engines: HOOMD-blue, LAMMPS, MolTwister internal (moldyn)

Exploring the aggregation of amyloid-β 42 through Monte Carlo simulations

Coarse-grained Monte Carlo simulations are performed for a disordered protein, amyloid-β 42 to identify the interactions and understand the mechanism of its aggregation. A statistical potential is developed from a selected dataset of intrinsically disordered proteins, which accounts for the respective contributions of the bonded and non-bonded potentials. While, the bonded potential comprises the bond, bend, and dihedral constraints, the nonbonded interactions include van der Waals interactions, hydrogen bonds, and the two-body potential. The two-body potential captures the features of both hydrophobic and electrostatic interactions that brings the chains at a contact distance, while the repulsive van der Waals interactions prevent them from a collapse. Increased two-body hydrophobic interactions facilitate the formation of amorphous aggregates rather than the fibrillar ones. The formation of aggregates is validated from the interchain distances, and the total energy of the system. The aggregate is structurally characterized by the root-mean-square deviation, root-mean-square fluctuation and the radius of gyration. The aggregates are characterized by a decrease in SASA, an increase in the non-local interactions and a distinct free energy minimum relative to that of the monomeric state of amyloid-β 42. The hydrophobic residues help in nucleation, while the charged residues help in oligomerization and aggregation.

Microphase behaviors and shear moduli of double-network gels: The effect of crosslinking constraints and chain uncrossability.

By performing coarse-grained molecular dynamics simulations, we study the effect of crosslinking and chain uncrossability on the microphase behaviors and mechanical properties of the double-network gels. The double-network systems can be viewed as two separate networks interpenetrating each other uniformly, and the crosslinks in each network are generated, forming a regular cubic lattice. The chain uncrossability is confirmed by appropriately choosing the bonded and nonbonded interaction potentials. Our simulations reveal a close relation between the phase and mechanical properties of the double-network systems and their network topological structures. Depending on the lattice size and the solvent affinity, we have observed two different microphases: one is the aggregation of solvophobic beads around the crosslinking points, which leads to locally polymer-rich domains, and the other is the bunching of polymer strands, which thickens the network edges and thus changes the network periodicity. The former is a representation of the interfacial effect, while the latter is determined by the chain uncrossability. The coalescence of network edges is demonstrated to be responsible for the large relative increase in the shear modulus. Compressing and stretching induced phase transitions are observed in the current double-network systems, and the sharp discontinuous change in the stress that appears at the transition point is found to be related to the bunching or debunching of the network edges. The results suggest that the regulation of network edges has a strong influence on the network mechanical properties.

Correlations in Hard- and Soft-Core Generic Polymer Models

Generic polymer models capturing the chain connectivity and the non-bonded excluded-volume interactions between polymer segments can be classified into hard- and soft-core models depending on their non-bonded pair potential. Here we compared the correlation effects on the structural and thermodynamic properties of the hard- and soft-core models given by the polymer reference interaction site model (PRISM) theory, and found different behaviors of the soft-core models at large invariant degree of polymerization (IDP) depending on how IDP is varied. We also proposed an efficient numerical approach, which enables us to accurately solve the PRISM theory for chain lengths as large as 106.

Molecular dynamics simulation insight into topological structure dependence of self-healing polymer nanocomposites.

Polymer nanocomposites (PNCs), which exhibit excellent mechanical properties through the incorporation of fillers into polymers, have been extensively studied to achieve enhanced self-healing capability for their next-generation development. However, there is still a lack of investigation into the influence of the topological structures of nanoparticles (NPs) on the self-healing capability of PNCs. In this study, we utilized coarse-grained molecular dynamics simulations (CGMDs) to construct a series of PNC systems composed of NPs with different topological structures, including Linear, Ring, and Cross topologies. We employed non-bonding interaction potentials to examine the interactions between the polymer and NPs, and varied the parameters to simulate different functional groups. Our results indicate that the stress-strain curves and the rate of performance loss validate that the Linear structure is the optimal topology for mechanical reinforcement and self-healing properties. By analyzing the stress heat map during stretching, we observed that the Linear structure NPs experience significant stress, allowing the matrix chains to dominate in small recoverable deformations during stretching. It can be speculated that NPs oriented in the direction of extrusion are more effective than others in enhancing performance. Overall, this work provides valuable theoretical guidance and a novel strategy for designing and manipulating high-performance, self-healing PNCs.

Effects of Residual Composition and Distribution on the Structural Characteristics of the Protein.

The effect of ratio and consecutive number of hydrophobic residues in the repeating unit of protein chains was investigated by MD simulation. The modified off-lattice HNP model was applied in this study. The protein chains constituted by different HNP ratios or different numbers of consecutively hydrophobic residues with the same chain length were simulated under a broad temperature range. We concluded that the proteins with higher ratio or larger number of sequentially hydrophobic residues present more orientated and compact structure under a certain low temperature. It is attributed to the lower non-bonded potential energy between H-H residual pairs, especially more hydrophobic residues in a procession among the protein chain. Considering the microscopic structure of the protein, more residue contacts are achieved with the proteins with higher ratios and sequential H residues under the low temperature. Meanwhile, with the ratio and consecutive number of H residues increasing, the distribution of stem length showed a transition from exponential decline to unimodal and even multiple peaks, indicating the specific ordered structure formed. These results provide an insight into 3D structural properties of proteins from their residue sequences, which has a primary structure at molecular level and, ultimately, a practical possibility of applying in biotechnological applications.

Coarse-Grained Model Incorporating Short- and Long-Range Effective Potentials for the Fast Simulation of Micelle Formation in Solutions of Ionic Surfactants.

A coarse-grained model comprising short- and long-range effective potentials, parametrized with the iterative Boltzmann inversion (IBI) method, is presented for capturing micelle formation in aqueous solutions of ionic surfactants using as a model system sodium dodecyl sulfate (SDS). In the coarse-grained (CG) model, each SDS molecule is represented as a sequence of four beads while each water molecule is modeled as a single bead. The proposed CG scheme involves ten potential energy functions: four of them describe bonded interactions and control the distribution functions of intramolecular degrees of freedom (bond lengths, valence angles, and dihedrals) along an SDS molecule while the other six account for intermolecular interactions between pairs of SDS and water beads and control the radial distribution functions. The nonbonded effective potentials between coarse-grained SDS molecules extend up to about 12 nm and capture structural and morphological features of the micellar solution both at short and long distances. The long-range component of these potentials, in particular, captures correlations between surfactant molecules belonging to different micelles and is essential to describe ordering associated with micelle formation. A new strategy is introduced for determining the effective potentials through IBI by using information (target distribution functions) extracted from independent atomistic simulations of a micellar reference system (a salt-free SDS solution at total surfactant concentration cT equal to 103 mM, temperature T equal to 300 K, and pressure P equal to 1 atm) obtained through a multiscale approach described in an earlier study. It employs several optimization steps for bonded and nonbonded interactions and a gradual parametrization of the short- and long-range components of the latter, followed by reparametrization of the bonded ones. The proposed CG model can reproduce remarkably accurately the microstructure and morphology of the reference system within only a few hours of computational time. It is therefore very promising for future studies of structural and morphological behavior of various liquid surfactant formulations.

Modeling Alkyl Aromatic Hydrocarbons with Dissipative Particle Dynamics.

Building on previous work studying alkanes, we develop a dissipative particle dynamics (DPD) model to capture the behavior of the alkyl aromatic hydrocarbon family under ambient conditions of 298 K and 1 atmosphere. Such materials are of significant worldwide industrial importance in applications such as solvents, chemical intermediates, surfactants, lubricating oils, hydraulic fluids, and greases. We model both liquids and waxy solids for molecules up to 36 carbons in size and demonstrate that we can correctly capture both the freezing transition and liquid-phase densities in pure substances and mixtures. We also demonstrate the importance of including specialized bead types into the DPD model (rather than solely relying on generic bead types) to capture specific local geometrical constructs such as the benzene ring found in the benzyl chemical group; this can be thought of as representing subtle real-world many-body effects via customized pairwise non-bonded potentials.

Slip-Spring Hybrid Particle-Field Molecular Dynamics for Coarse-Graining Branched Polymer Melts: Polystyrene Melts as an Example.

The topology of chains significantly modifies the dynamical properties of polymer melts. Here, we extend a recently developed efficient simulation method, namely the slip-spring hybrid particle-field (SS-hPF) model, to study the structural and dynamical properties of branched polymer melts over large spatial-temporal scales. In the coarse-grained SS-hPF simulation of polymers, the bonded potentials are derived by iterative Boltzmann inversion from the underlying fine-grained model. The nonbonded potentials are computed from a density functional field instead of pairwise interactions used in standard molecular dynamics simulations, which increases the computational efficiency by a factor of 10-20. The entangled dynamics is lost due to the soft-core nature of density functional field interactions. It is recovered by a multichain slip-spring model that is rigorously parametrized from existing experimental or simulation data. To quantitatively predict the relaxation and diffusion of branched polymers, which are dominated by arm retraction rather than chain reptation, the slip-spring algorithm is augmented to improve the polymer dynamics near the branch point. Multiple dynamical observables, e.g., diffusion coefficients, arm relaxations, and tube survival probabilities, are characterized in an example coarse-grained model of symmetric and asymmetric star-shaped polystyrene melts. Consistent dynamical behaviors are identified and compared with theoretical predictions. With a single rescaling factor, the prediction of diffusion coefficients agrees well with the available experimental measurements. In this work, an efficient approach is provided to build chemistry-specific coarse-grained models for predicting the dynamics of branched polymers.

Force Field Parameterization of Actinyl Molecular Cations Using the 12-6-4 Model.

In this work, a set of 12-6-4 force fields (FFs) parameters were developed for the actinyl molecular cations, AnO2n+ (n = 1, 2), from uranium to plutonium for classical molecular dynamics (MD) for four water models: TIP3P, SPC/E, OPC3, and TIP4Pew. Such a non-bonded potential model taking into account the induced dipole between the metallic center and the surrounding molecules has shown better performances for various cations than the classic 12-6 non-bonded potentials. The parametrization method proposed elsewhere for metallic cations has been extended to these molecular cations. In contrast to the actinyl 12-6 FFs from the literature, the new models reproduce correctly both solvation and thermodynamic properties, thanks to the inclusion of the induced dipole term (C4). The transferability of such force fields was assessed by performing MD simulations of carbonato actinyl species, which are highly implicated in actinide migration or actinide extraction from seawater. A highly satisfying agreement was found when comparing the EXAFS signals computed from our MD simulation to the experimental ones. The set of FFs developed here opens new possibilities for the study of actinide chemistry.

Implementation of residue-level coarse-grained models in GENESIS for large-scale molecular dynamics simulations.

Residue-level coarse-grained (CG) models have become one of the most popular tools in biomolecular simulations in the trade-off between modeling accuracy and computational efficiency. To investigate large-scale biological phenomena in molecular dynamics (MD) simulations with CG models, unified treatments of proteins and nucleic acids, as well as efficient parallel computations, are indispensable. In the GENESIS MD software, we implement several residue-level CG models, covering structure-based and context-based potentials for both well-folded biomolecules and intrinsically disordered regions. An amino acid residue in protein is represented as a single CG particle centered at the Cα atom position, while a nucleotide in RNA or DNA is modeled with three beads. Then, a single CG particle represents around ten heavy atoms in both proteins and nucleic acids. The input data in CG MD simulations are treated as GROMACS-style input files generated from a newly developed toolbox, GENESIS-CG-tool. To optimize the performance in CG MD simulations, we utilize multiple neighbor lists, each of which is attached to a different nonbonded interaction potential in the cell-linked list method. We found that random number generations for Gaussian distributions in the Langevin thermostat are one of the bottlenecks in CG MD simulations. Therefore, we parallelize the computations with message-passing-interface (MPI) to improve the performance on PC clusters or supercomputers. We simulate Herpes simplex virus (HSV) type 2 B-capsid and chromatin models containing more than 1,000 nucleosomes in GENESIS as examples of large-scale biomolecular simulations with residue-level CG models. This framework extends accessible spatial and temporal scales by multi-scale simulations to study biologically relevant phenomena, such as genome-scale chromatin folding or phase-separated membrane-less condensations.

Mesoscopic Modeling of a Highly-Ordered Sanidic Polymer Mesophase and Comparison With Experimental Data.

Board-shaped polymers form sanidic mesophases: assemblies of parallel lamellae of stacked polymer backbones separated by disordered side chains. Sanidics vary significantly with respect to polymer order inside their lamellae, making them “stepping stones” toward the crystalline state. Therefore, they are potentially interesting for studying crystallization and technological applications. Building on earlier mesoscopic models of the most disordered sanidics Σd, we focus on the other extreme, near-crystalline order, and develop a generic model that captures a highly ordered Σr mesophase. Polymers are described by generic hindered-rotation chains. Anisotropic nonbonded potentials, with strengths comparable to the thermal energy, mimic board-like monomer shapes. Lamellae equilibrated with Monte Carlo simulations, for a broad range of model parameters, have intralamellar order typical for Σr mesophases: periodically stacked polymers that are mutually registered along their backbones. Our mesophase shows registration on both monomer and chain levels. We calculate scattering patterns and compare with data published for highly ordered sanidic mesophases of two different polymers: polyesters and polypeptoids. Most of the generic structural features that were identified in these experiments are present in our model. However, our mesophase has correlations between chains located in different lamellae and is therefore closer to the crystalline state than the experimental samples.

Role of mapping schemes on dynamical and mechanical properties of coarse-grained models of cis-1,4-polyisoprene

We study and compare four coarse-grained models of cis-1,4-polyisoprene distinguished by mapping schemes for locating superatoms. First, coarse-grained potentials are obtained by iterative Boltzmann inversion method for the polymer melt. For all the coarse-grained models considered, time-scale factors based on translational and rotational motion were found to be different. However, coarse-grained potentials were unable to reproduce stress–strain behaviour of the underlying detailed model due to weak attractive nature of the nonbonded part of the coarse-grained potential. Consequently, the nonbonded potentials were optimized using particle swarm optimization to match the tensile behaviour of a detailed model of the polymer below glass transition temperature. While the modified potentials seemed to better predict the mechanical behaviour, the ability to accurately predict simultaneously the structural distributions also, depends on the mapping scheme.

Short-Range Imbalances in the AMBER Lennard-Jones Potential for (Deoxy)Ribose···Nucleobase Lone-Pair···π Contacts in Nucleic Acids.

The lone-pair···π (lp···π) (deoxy)ribose···nucleobase stacking is a recurring interaction in Z-DNA and RNAs that is characterized by sub-van der Waals lp···π contacts (<3.0 Å). It is a part of the structural signature of CpG Z-step motifs in Z-DNA and r(UNCG) tetraloops that are known to behave poorly in molecular dynamics (MD) simulations. Although the exact origin of the MD simulation issues remains unclear, a significant part of the problem might be due to an imbalanced description of nonbonded interactions, including the characteristic lp···π stacking. To gain insights into the links between lp···π stacking and MD, we present an in-depth comparison between accurate large-basis-set double-hybrid Kohn-Sham density functional theory calculations DSD-BLYP-D3/ma-def2-QZVPP (DHDF-D3) and data obtained with the nonbonded potential of the AMBER force field (AFF) for NpN Z-steps (N = G, A, C, and U). Among other differences, we found that the AFF overestimates the DHDF-D3 lp···π distances by ∼0.1-0.2 Å, while the deviation between the DHDF-D3 and AFF descriptions sharply increases in the short-range region of the interaction. Based on atom-in-molecule polarizabilities and symmetry-adapted perturbation theory analysis, we inferred that the DHDF-D3 versus AFF differences partly originate in identical nucleobase carbon atom Lennard-Jones (LJ) parameters despite the presence/absence of connected electron-withdrawing groups that lead to different effective volumes or vdW radii. Thus, to precisely model the very short CpG lp···π contact distances, we recommend revision of the nucleobase atom LJ parameters. Additionally, we suggest that the large discrepancy between DHDF-D3 and AFF short-range repulsive part of the interaction energy potential may significantly contribute to the poor performances of MD simulations of nucleic acid systems containing Z-steps. Understanding where, and if possible why, the point-charge-type effective potentials reach their limits is vital for developing next-generation FFs and for addressing specific issues in contemporary MD simulations.

SMOG 2 and OpenSMOG: Extending the limits of structure-based models.

Applying simulations with structure-based models has proven to be an effective strategy for investigating the factors that control biomolecular dynamics. The common element of these models is that some (or all) of the intra/inter-molecular interactions are explicitly defined to stabilize an experimentally determined structure. To facilitate the development and application of this broad class of models, we previously released the SMOG 2 software package. This suite allows one to easily customize and distribute structure-based (i.e., SMOG) models for any type of polymer-ligand system. The force fields generated by SMOG 2 may then be used to perform simulations in highly optimized MD packages, such as Gromacs, NAMD, LAMMPS, and OpenMM. Here, we describe extensions to the software and demonstrate the capabilities of the most recent version (SMOG v2.4.2). Changes include new tools that aid user-defined customization of force fields, as well as an interface with the OpenMM simulation libraries (OpenSMOG v1.1.0). The OpenSMOG module allows for arbitrary user-defined contact potentials and non-bonded potentials to be employed in SMOG models, without source-code modifications. To illustrate the utility of these advances, we present applications to systems with millions of atoms, long polymers and explicit ions, as well as models that include non-structure-based (e.g., AMBER-based) energetic terms. Examples include large-scale rearrangements of the SARS-CoV-2 Spike protein, the HIV-1 capsid with explicit ions, and crystallographic lattices of ribosomes and proteins. In summary, SMOG 2 and OpenSMOG provide robust support for researchers who seek to develop and apply structure-based models to large and/or intricate biomolecular systems.