GLIMPS: A Machine Learning Approach to Resolution Transformation for Multiscale Modeling

Chemically Transferable Electronic Coarse Graining for Polythiophenes

Recent advances in machine-learning-based electronic coarse graining (ECG) methods have demonstrated the potential to enable electronic predictions in soft materials at mesoscopic length scales. However, previous ECG models have yet to confront the issue of chemical transferability. In this study, we develop chemically transferable ECG models for polythiophenes using graph neural networks. Our models are trained on a data set that samples over the conformational space of random polythiophene sequences generated with 15 different monomer chemistries and three different degrees of polymerization. We systematically explore the impact of coarse-grained representation on ECG accuracy, highlighting the significance of preserving the C-β coordinates in thiophene. We also find that integrating unique polymer sequences into training enhances the model performance more efficiently than augmenting conformational sampling for sequences already in the training data set. Moreover, our ECG models, developed initially for one property and one level of quantum chemical theory, can be efficiently transferred to related properties and higher levels of theory with minimal additional data. The chemically transferable ECG model introduced in this work will serve as a foundation model for new classes of chemically transferable ECG predictions across chemical space.

Analysis of mapping atomic models to coarse-grained resolution

Low-resolution coarse-grained (CG) models provide significant computational and conceptual advantages for simulating soft materials. However, the properties of CG models depend quite sensitively upon the mapping, M, that maps each atomic configuration, r, to a CG configuration, R. In particular, M determines how the configurational information of the atomic model is partitioned between the mapped ensemble of CG configurations and the lost ensemble of atomic configurations that map to each R. In this work, we investigate how the mapping partitions the atomic configuration space into CG and intra-site components. We demonstrate that the corresponding coordinate transformation introduces a nontrivial Jacobian factor. This Jacobian factor defines a labeling entropy that corresponds to the uncertainty in the atoms that are associated with each CG site. Consequently, the labeling entropy effectively transfers configurational information from the lost ensemble into the mapped ensemble. Moreover, our analysis highlights the possibility of resonant mappings that separate the atomic potential into CG and intra-site contributions. We numerically illustrate these considerations with a Gaussian network model for the equilibrium fluctuations of actin. We demonstrate that the spectral quality, Q, provides a simple metric for identifying high quality representations for actin. Conversely, we find that neither maximizing nor minimizing the information content of the mapped ensemble results in high quality representations. However, if one accounts for the labeling uncertainty, Q(M) correlates quite well with the adjusted configurational information loss, Îmap(M), that results from the mapping.

ezAlign: A Tool for Converting Coarse-Grained Molecular Dynamics Structures to Atomistic Resolution for Multiscale Modeling

Soft condensed matter is challenging to study due to the vast time and length scales that are necessary to accurately represent complex systems and capture their underlying physics. Multiscale simulations are necessary to study processes that have disparate time and/or length scales, which abound throughout biology and other complex systems. Herein we present ezAlign, an open-source software for converting coarse-grained molecular dynamics structures to atomistic representation, allowing multiscale modeling of biomolecular systems. The ezAlign v1.1 software package is publicly available for download at github.com/LLNL/ezAlign. Its underlying methodology is based on a simple alignment of an atomistic template molecule, followed by position-restraint energy minimization, which forces the atomistic molecule to adopt a conformation consistent with the coarse-grained molecule. The molecules are then combined, solvated, minimized, and equilibrated with position restraints. Validation of the process was conducted on a pure POPC membrane and compared with other popular methods to construct atomistic membranes. Additional examples, including surfactant self-assembly, membrane proteins, and more complex bacterial and human plasma membrane models, are also presented. By providing these examples, parameter files, code, and an easy-to-follow recipe to add new molecules, this work will aid future multiscale modeling efforts.

CGeNArate: a sequence-dependent coarse-grained model of DNA for accurate atomistic MD simulations of kb-long duplexes

We present CGeNArate, a new model for molecular dynamics simulations of very long segments of B-DNA in the context of biotechnological or chromatin studies. The developed method uses a coarse-grained Hamiltonian with trajectories that are back-mapped to the atomistic resolution level with extreme accuracy by means of Machine Learning Approaches. The method is sequence-dependent and reproduces very well not only local, but also global physical properties of DNA. The efficiency of the method allows us to recover with a reduced computational effort high-quality atomic-resolution ensembles of segments containing many kilobases of DNA, entering into the gene range or even the entire DNA of certain cellular organelles.

Building complex membranes with Martini 3

The Martini model is a popular force field for coarse-grained simulations. Membranes have always been at the center of its development, with the latest version, Martini 3, showing great promise in capturing more and more realistic behavior. In this chapter we provide a step-by-step tutorial on how to construct starting configurations, run initial simulations and perform dedicated analysis for membrane-based systems of increasing complexity, including leaflet asymmetry, curvature gradients and embedding of membrane proteins.

Improved Protein Dynamics and Hydration in the Martini3 Coarse-Grain Model

The Martini coarse-grain force-field has emerged as an important framework to probe cellular processes at experimentally relevant time- and length-scales. However, the recently developed version, the Martini3 force-field with the implemented Go̅ model (Martini3Go̅), as well as previous variants of the Martini model have not been benchmarked and rigorously tested for globular proteins. In this study, we consider three globular proteins, ubiquitin, lysozyme, and cofilin, and compare protein dynamics and hydration with observables from experiments and all-atom simulations. We show that the Martini3Go̅ model is able to accurately model the structural and dynamic features of small globular proteins. Overall, the structural integrity of the proteins is maintained, as validated by contact maps, radii of gyration (Rg), and SAXS profiles. The chemical shifts predicted from the ensemble sampled in the simulations are consistent with the experimental data. Further, a good match is observed in the protein-water interaction energetics, and the hydration levels of the residues are similar to atomistic simulations. However, the protein-water interaction dynamics is not accurately represented and appears to depend on the protein structural complexity, residue specificity, and water dynamics. Our work is a step toward testing and assessing the Martini3Go̅ model and provides insights into future efforts to refine Martini models with improved solvation effects and better correspondence to the underlying all-atom systems.

Surveying the energy landscape of coarse-grained mappings

Simulations of soft materials often adopt low-resolution coarse-grained (CG) models. However, the CG representation is not unique and its impact upon simulated properties is poorly understood. In this work, we investigate the space of CG representations for ubiquitin, which is a typical globular protein with 72 amino acids. We employ Monte Carlo methods to ergodically sample this space and to characterize its landscape. By adopting the Gaussian network model as an analytically tractable atomistic model for equilibrium fluctuations, we exactly assess the intrinsic quality of each CG representation without introducing any approximations in sampling configurations or in modeling interactions. We focus on two metrics, the spectral quality and the information content, that quantify the extent to which the CG representation preserves low-frequency, large-amplitude motions and configurational information, respectively. The spectral quality and information content are weakly correlated among high-resolution representations but become strongly anticorrelated among low-resolution representations. Representations with maximal spectral quality appear consistent with physical intuition, while low-resolution representations with maximal information content do not. Interestingly, quenching studies indicate that the energy landscape of mapping space is very smooth and highly connected. Moreover, our study suggests a critical resolution below which a "phase transition" qualitatively distinguishes good and bad representations.

Distilling coarse-grained representations of molecular electronic structure with continuously gated message passing

Bottom-up methods for coarse-grained (CG) molecular modeling are critically needed to establish rigorous links between atomistic reference data and reduced molecular representations. For a target molecule, the ideal reduced CG representation is a function of both the conformational ensemble of the system and the target physical observable(s) to be reproduced at the CG resolution. However, there is an absence of algorithms for selecting CG representations of molecules from which complex properties, including molecular electronic structure, can be accurately modeled. We introduce continuously gated message passing (CGMP), a graph neural network (GNN) method for atomically decomposing molecular electronic structure sampled over conformational ensembles. CGMP integrates 3D-invariant GNNs and a novel gated message passing system to continuously reduce the atomic degrees of freedom accessible for electronic predictions, resulting in a one-shot importance ranking of atoms contributing to a target molecular property. Moreover, CGMP provides the first approach by which to quantify the degeneracy of "good" CG representations conditioned on specific prediction targets, facilitating the development of more transferable CG representations. We further show how CGMP can be used to highlight multiatom correlations, illuminating a path to developing CG electronic Hamiltonians in terms of interpretable collective variables for arbitrarily complex molecules.

Martini 3 Force Field Parameters for Protein Lipidation Post-Translational Modifications

Protein lipidations are vital co/post-translational modifications that tether lipid tails to specific protein amino acids, allowing them to anchor to biological membranes, switch their subcellular localization, and modulate association with other proteins. Such lipidations are thus crucial for multiple biological processes including signal transduction, protein trafficking, and membrane localization and are implicated in various diseases as well. Examples of lipid-anchored proteins include the Ras family of proteins that undergo farnesylation; actin and gelsolin that are myristoylated; phospholipase D that is palmitoylated; glycosylphosphatidylinositol-anchored proteins; and others. Here, we develop parameters for cysteine-targeting farnesylation, geranylgeranylation, and palmitoylation, as well as glycine-targeting myristoylation for the latest version of the Martini 3 coarse-grained force field. The parameters are developed using the CHARMM36m all-atom force field parameters as reference. The behavior of the coarse-grained models is consistent with that of the all-atom force field for all lipidations and reproduces key dynamical and structural features of lipid-anchored peptides, such as the solvent-accessible surface area, bilayer penetration depth, and representative conformations of the anchors. The parameters are also validated in simulations of the lipid-anchored peripheral membrane proteins Rheb and Arf1, after comparison with independent all-atom simulations. The parameters, along with mapping schemes for the popular martinize2 tool, are available for download at 10.5281/zenodo.7849262 and also as supporting information.

Backmapping with Mapping and Isomeric Information

I present a powerful and flexible backmapping tool named Multiscale Simulation Tool (mstool) that converts a coarse-grained (CG) system into all-atom (AA) resolution and only requires AA to CG mapping and isomeric information (cis/trans/dihedral/chiral). The backmapping procedure includes two simple steps: (a) AA atoms are randomly placed near the corresponding CG beads according to the provided mapping scheme. (b) Energy minimization is performed with two modifications in the AA force field (FF). First, nonbonded interactions are replaced with cosine functions to ensure the numerical stability. Second, additional torsions are imposed to maintain the molecules' isomeric properties. To test the simplicity and robustness of the tool, I backmapped multiple membrane and protein CG structures into AA resolution, including a four-bead CG lipid model (resolution increased by a factor of 34) without using intermediate resolution. The tool is freely available at github.com/ksy141/mstool.

Machine Learning-Driven Multiscale Modeling: Bridging the Scales with a Next-Generation Simulation Infrastructure

Interdependence across time and length scales is common in biology, where atomic interactions can impact larger-scale phenomenon. Such dependence is especially true for a well-known cancer signaling pathway, where the membrane-bound RAS protein binds an effector protein called RAF. To capture the driving forces that bring RAS and RAF (represented as two domains, RBD and CRD) together on the plasma membrane, simulations with the ability to calculate atomic detail while having long time and large length- scales are needed. The Multiscale Machine-Learned Modeling Infrastructure (MuMMI) is able to resolve RAS/RAF protein-membrane interactions that identify specific lipid-protein fingerprints that enhance protein orientations viable for effector binding. MuMMI is a fully automated, ensemble-based multiscale approach connecting three resolution scales: (1) the coarsest scale is a continuum model able to simulate milliseconds of time for a 1 μm2 membrane, (2) the middle scale is a coarse-grained (CG) Martini bead model to explore protein-lipid interactions, and (3) the finest scale is an all-atom (AA) model capturing specific interactions between lipids and proteins. MuMMI dynamically couples adjacent scales in a pairwise manner using machine learning (ML). The dynamic coupling allows for better sampling of the refined scale from the adjacent coarse scale (forward) and on-the-fly feedback to improve the fidelity of the coarser scale from the adjacent refined scale (backward). MuMMI operates efficiently at any scale, from a few compute nodes to the largest supercomputers in the world, and is generalizable to simulate different systems. As computing resources continue to increase and multiscale methods continue to advance, fully automated multiscale simulations (like MuMMI) will be commonly used to address complex science questions.

Integrating Machine Learning in the Coarse-Grained Molecular Simulation of Polymers

Machine learning (ML) is having an increasing impact on the physical sciences, engineering, and technology and its integration into molecular simulation frameworks holds great potential to expand their scope of applicability to complex materials and facilitate fundamental knowledge and reliable property predictions, contributing to the development of efficient materials design routes. The application of ML in materials informatics in general, and polymer informatics in particular, has led to interesting results, however great untapped potential lies in the integration of ML techniques into the multiscale molecular simulation methods for the study of macromolecular systems, specifically in the context of Coarse Grained (CG) simulations. In this Perspective, we aim at presenting the pioneering recent research efforts in this direction and discussing how these new ML-based techniques can contribute to critical aspects of the development of multiscale molecular simulation methods for bulk complex chemical systems, especially polymers. Prerequisites for the implementation of such ML-integrated methods and open challenges that need to be met toward the development of general systematic ML-based coarse graining schemes for polymers are discussed.

Molecular dynamics simulation of an entire cell

The ultimate microscope, directed at a cell, would reveal the dynamics of all the cell's components with atomic resolution. In contrast to their real-world counterparts, computational microscopes are currently on the brink of meeting this challenge. In this perspective, we show how an integrative approach can be employed to model an entire cell, the minimal cell, JCVI-syn3A, at full complexity. This step opens the way to interrogate the cell's spatio-temporal evolution with molecular dynamics simulations, an approach that can be extended to other cell types in the near future.

Deep convolutional neural networks for generating atomistic configurations of multi-component macromolecules from coarse-grained models

Despite the modern advances in the available computational resources, the length and time scales of the physical systems that can be studied in full atomic detail, via molecular simulations, are still limited. To overcome such limitations, coarse-grained (CG) models have been developed to reduce the dimensionality of the physical system under study. However, to study such systems at the atomic level, it is necessary to re-introduce the atomistic details into the CG description. Such an ill-posed mathematical problem is typically treated via numerical algorithms, which need to balance accuracy, efficiency, and general applicability. Here, we introduce an efficient and versatile method for backmapping multi-component CG macromolecules of arbitrary microstructures. By utilizing deep learning algorithms, we train a convolutional neural network to learn structural correlations between polymer configurations at the atomistic and their corresponding CG descriptions, obtained from atomistic simulations. The trained model is then utilized to get predictions of atomistic structures from input CG configurations. As an illustrative example, we apply the convolutional neural network to polybutadiene copolymers of various microstructures, in which each monomer microstructure (i.e., cis-1,4, trans-1,4, and vinyl-1,2) is represented as a different CG particle type. The proposed methodology is transferable over molecular weight and various microstructures. Moreover, starting from a specific single CG configuration with a given microstructure, we show that by modifying its chemistry (i.e., CG particle types), we are able to obtain a set of well equilibrated polymer configurations of different microstructures (chemistry) than the one of the original CG configuration.

Bottom-up Coarse-Graining: Principles and Perspectives

Large-scale computational molecular models provide scientists a means to investigate the effect of microscopic details on emergent mesoscopic behavior. Elucidating the relationship between variations on the molecular scale and macroscopic observable properties facilitates an understanding of the molecular interactions driving the properties of real world materials and complex systems (e.g., those found in biology, chemistry, and materials science). As a result, discovering an explicit, systematic connection between microscopic nature and emergent mesoscopic behavior is a fundamental goal for this type of investigation. The molecular forces critical to driving the behavior of complex heterogeneous systems are often unclear. More problematically, simulations of representative model systems are often prohibitively expensive from both spatial and temporal perspectives, impeding straightforward investigations over possible hypotheses characterizing molecular behavior. While the reduction in resolution of a study, such as moving from an atomistic simulation to that of the resolution of large coarse-grained (CG) groups of atoms, can partially ameliorate the cost of individual simulations, the relationship between the proposed microscopic details and this intermediate resolution is nontrivial and presents new obstacles to study. Small portions of these complex systems can be realistically simulated. Alone, these smaller simulations likely do not provide insight into collectively emergent behavior. However, by proposing that the driving forces in both smaller and larger systems (containing many related copies of the smaller system) have an explicit connection, systematic bottom-up CG techniques can be used to transfer CG hypotheses discovered using a smaller scale system to a larger system of primary interest. The proposed connection between different CG systems is prescribed by (i) the CG representation (mapping) and (ii) the functional form and parameters used to represent the CG energetics, which approximate potentials of mean force (PMFs). As a result, the design of CG methods that facilitate a variety of physically relevant representations, approximations, and force fields is critical to moving the frontier of systematic CG forward. Crucially, the proposed connection between the system used for parametrization and the system of interest is orthogonal to the optimization used to approximate the potential of mean force present in all systematic CG methods. The empirical efficacy of machine learning techniques on a variety of tasks provides strong motivation to consider these approaches for approximating the PMF and analyzing these approximations.

Asynchronous Reciprocal Coupling of Martini 2.2 Coarse-Grained and CHARMM36 All-Atom Simulations in an Automated Multiscale Framework

The appeal of multiscale modeling approaches is predicated on the promise of combinatorial synergy. However, this promise can only be realized when distinct scales are combined with reciprocal consistency. Here, we consider multiscale molecular dynamics (MD) simulations that combine the accuracy and macromolecular flexibility accessible to fixed-charge all-atom (AA) representations with the sampling speed accessible to reductive, coarse-grained (CG) representations. AA-to-CG conversions are relatively straightforward because deterministic routines with unique outcomes are achievable. Conversely, CG-to-AA conversions have many solutions due to a surge in the number of degrees of freedom. While automated tools for biomolecular CG-to-AA transformation exist, we find that one popular option, called Backward, is prone to stochastic failure and the AA models that it does generate frequently have compromised protein structure and incorrect stereochemistry. Although these shortcomings can likely be circumvented by human intervention in isolated instances, automated multiscale coupling requires reliable and robust scale conversion. Here, we detail an extension to Multiscale Machine-learned Modeling Infrastructure (MuMMI), including an improved CG-to-AA conversion tool called sinceCG. This tool is reliable (∼98% weakly correlated repeat success rate), automatable (no unrecoverable hangs), and yields AA models that generally preserve protein secondary structure and maintain correct stereochemistry. We describe how the MuMMI framework identifies CG system configurations of interest, converts them to AA representations, and simulates them at the AA scale while on-the-fly analyses provide feedback to update CG parameters. Application to systems containing the peripheral membrane protein RAS and proximal components of RAF kinase on complex eight-component lipid bilayers with ∼1.5 million atoms is discussed in the context of MuMMI.