PSO-Assisted Development of New Transferable Coarse-Grained Water Models

Can Coarse-Grained Molecular Dynamics Simulations Predict Pharmaceutical Crystal Growth?

To investigate the ability of coarse-grained molecular dynamics simulations to predict the relative growth rates of crystal facets of pharmaceutical molecules, we apply two coarse-graining strategies to two drug molecules, phenytoin and carbamazepine. In the first method, we map an atomistic model to a MARTINI-level coarse-grained (CG) force field that uses 2 or 3 heavy atoms per bead. This is followed by applying Particle Swarm Optimization (PSO), a global optimum searching algorithm, to the CG Lennard-Jones intermolecular potentials to fit the radial distribution functions of both the crystalline and melt structures. In the second, a coarser-grained method, we map 5 or more heavy atoms into one bead with the help of the Iterative Boltzmann Inversion (IBI) method to derive a tabulated longer-range force field (FF). Simulations using the FF's derived from both strategies were able to stabilize the crystal in the correct structure and to predict crystal growth from the melt with modest computational resources. We evaluate the advantages and limitations of both methods and compare the relative growth rates of various facets of both drug crystals with those predicted by the Bravais-Friedel-Donnay-Harker (BFDH) and attachment energy (AE) theories. While all methods, except for the simulations conducted with the coarser-grained IBI-generated model, produced similarly good results for phenytoin, the finer-grained PSO-generated FF using MARTINI mapping rules outperformed the other methods in its prediction of the facet growth rates and resulting crystalline morphology for carbamazepine.

Evaluation of Sampling Algorithms Used for Bayesian Uncertainty Quantification of Molecular Dynamics Force Fields

New Bayesian parameter estimation methods have the capability to enable more physically realistic and reliable molecular dynamics (MD) simulations by providing accurate estimates of uncertainties of force-field (FF) parameters and associated properties. However, the choice of which Bayesian parameter estimation algorithm to use has not been widely investigated, despite its impact on the effective exploration of parameter space. Here, using a case example of the Embedded Atom Method (EAM) FF parameters, we investigated the ramifications of several of the algorithm choices. We found that Ensemble Slice Sampling (ESS) and Affine-Invariant Ensemble Sampling (AIES) demonstrate a new level of superior performance, culminating in more accurate parameter and property estimations with tighter uncertainty bounds, compared to traditional methods such as Metropolis-Hastings (MH), Gradient Search (GS), and Uniform Random Sampler (URS). We demonstrate that Bayesian Uncertainty Quantification with ESS and AIES leads to significantly more accurate and reliable predictions of the FF parameters and properties. The results suggest that ESS and AIES should be used to obtain more accurate parameter and uncertainty estimations while providing deeper physical insights.

CGCompiler: Automated Coarse-Grained Molecule Parametrization via Noise-Resistant Mixed-Variable Optimization

Coarse-grained force fields (CG FFs) such as the Martini model entail a predefined, fixed set of Lennard-Jones parameters (building blocks) to model virtually all possible nonbonded interactions between chemically relevant molecules. Owing to its universality and transferability, the building-block coarse-grained approach has gained tremendous popularity over the past decade. The parametrization of molecules can be highly complex and often involves the selection and fine-tuning of a large number of parameters (e.g., bead types and bond lengths) to optimally match multiple relevant targets simultaneously. The parametrization of a molecule within the building-block CG approach is a mixed-variable optimization problem: the nonbonded interactions are discrete variables, whereas the bonded interactions are continuous variables. Here, we pioneer the utility of mixed-variable particle swarm optimization in automatically parametrizing molecules within the Martini 3 coarse-grained force field by matching both structural (e.g., RDFs) as well as thermodynamic data (phase-transition temperatures). For the sake of demonstration, we parametrize the linker of the lipid sphingomyelin. The important advantage of our approach is that both bonded and nonbonded interactions are simultaneously optimized while conserving the search efficiency of vector guided particle swarm optimization (PSO) methods over other metaheuristic search methods such as genetic algorithms. In addition, we explore noise-mitigation strategies in matching the phase-transition temperatures of lipid membranes, where nucleation and concomitant hysteresis introduce a dominant noise term within the objective function. We propose that noise-resistant mixed-variable PSO methods can both improve and automate parametrization of molecules within building-block CG FFs, such as Martini.

Data Driven Discovery of MOFs for Hydrogen Gas Adsorption

Hydrogen gas (H2) is a clean and renewable energy source, but the lack of efficient and cost-effective storage materials is a challenge to its widespread use. Metal-organic frameworks (MOFs), a class of porous materials, have been extensively studied for H2 storage due to their tunable structural and chemical features. However, the large design space offered by MOFs makes it challenging to select or design appropriate MOFs with a high H2 storage capacity. To overcome these challenges, we present a data-driven computational approach that systematically designs new functionalized MOFs for H2 storage. In particular, we showcase the framework of a hybrid particle swarm optimization integrated genetic algorithm, grand canonical Monte Carlo (GCMC) simulations, and our in-house MOF structure generation code to design new MOFs with excellent H2 uptake. This automated, data driven framework adds appropriate functional groups to IRMOF-10 to improve its H2 adsorption capacity. A detailed analysis of the top selected MOFs, their adsorption isotherms, and MOF design rules to enhance H2 adsorption are presented. We found a functionalized IRMOF-10 with an enhanced H2 adsorption increased by ∼6 times compared to that of pure IRMOF-10 at 1 bar and 77 K. Furthermore, this study also utilizes machine learning and deep learning techniques to analyze a large data set of MOF structures and properties, in order to identify the key factors that influence hydrogen adsorption. The proof-of-concept that uses a machine learning/deep learning approach to predict hydrogen adsorption based on the identified structural and chemical properties of the MOF is demonstrated.

Extension of the SUGRES-1P Coarse-Grained Model of Polysaccharides to Heparin

Heparin is an unbranched periodic polysaccharide composed of negatively charged monomers and involved in key biological processes, including anticoagulation, angiogenesis, and inflammation. Its structure and dynamics have been studied extensively using experimental as well as theoretical approaches. The conventional approach of computational chemistry applied to the analysis of biomolecules is all-atom molecular dynamics, which captures the interactions of individual atoms by solving Newton's equation of motion. An alternative is molecular dynamics simulations using coarse-grained models of biomacromolecules, which offer a reduction of the representation and consequently enable us to extend the time and size scale of simulations by orders of magnitude. In this work, we extend the UNIfied COarse-gRaiNed (UNICORN) model of biological macromolecules developed in our laboratory to heparin. We carried out extensive tests to estimate the optimal weights of energy terms of the effective energy function as well as the optimal Debye-Hückel screening factor for electrostatic interactions. We applied the model to study unbound heparin molecules of polymerization degree ranging from 6 to 68 residues. We compare the obtained coarse-grained heparin conformations with models obtained from X-ray diffraction studies of heparin. The SUGRES-1P force field was able to accurately predict the general shape and global characteristics of heparin molecules.

Development, Validation, and Applications of Nonbonded Interaction Parameters between Coarse-Grained Amino Acid and Water Models

Interactions between amino acids and water play an important role in determining the stability and folding/unfolding, in aqueous solution, of many biological macromolecules, which affects their function. Thus, understanding the molecular-level interactions between water and amino acids is crucial to tune their function in aqueous solutions. Herein, we have developed nonbonded interaction parameters between the coarse-grained (CG) models of 20 amino acids and the one-site CG water model. The nonbonded parameters, represented using the 12-6 Lennard Jones (LJ) potential form, have been optimized using an artificial neural network (ANN)-assisted particle swarm optimization (PSO) (ANN-assisted PSO) method. All-atom (AA) molecular dynamics (MD) simulations of dipeptides in TIP3P water molecules were performed to calculate the Gibbs hydration free energies. The nonbonded force-field (FF) parameters between CG amino acids and the one-site CG water model were developed to accurately reproduce these energies. Furthermore, to test the transferability of these newly developed parameters, we calculated the hydration free energies of the analogues of the amino acid side chains, which showed good agreement with reported experimental data. Additionally, we show the applicability of these models by performing self-assembly simulations of peptide amphiphiles. Overall, these models are transferable and can be used to study the self-assembly of various biomaterials and biomolecules to develop a mechanistic understanding of these processes.

Structural and thermodynamic properties of bulk triglycerides and triglyceride/water mixtures reproduced using a polarizable coarse-grained model

Triglycerides (TGs) play important roles in renewable energies, food production, medicine, and metabolism in organisms. Here, we developed a novel coarse-grained (CG) force field (FF) for triglycerides to reproduce both the structural and thermodynamic properties of bulk TGs, TG/air interfaces, and TG/water mixtures using molecular dynamics (MD) simulations. We rigorously optimized the bonded and nonbonded force parameters between the CG beads of TGs and nonbonded force parameters between TG beads and polarizable CG water beads by employing an efficient meta-multilinear interpolation parameterization algorithm recently developed by us. This CG FF performs very well in reproducing the percolating network of the TG bulk phase self-assembled in water and a variety of molecular conformations predicted by all-atom MD simulations. More importantly, it also correctly reproduces multiple experimentally measurable macroscopic thermodynamic properties, including the density and surface tensions of both the TG/air and TG/water interfaces. This paves the way for studying more complicated systems involving TGs on a large scale.

Approaches and Perspective of Coarse-Grained Modeling and Simulation for Polymer-Nanoparticle Hybrid Systems

Molecular modeling and simulations have emerged as effective and indispensable tools to characterize polymeric systems. They provide fundamental and essential insights to design a product of the required properties and to improve the understanding of a phenomenon at the molecular level for a particular system. The polymer-nanoparticle hybrids are materials with outstanding properties and correspondingly large applications whose study has benefited from this new paradigm. However, despite the significant expansion of modern day computational powers, investigation of the long time and large length scale phenomenon in polymeric and polymer-nanoparticle systems is still a challenging task to complete through all-atom molecular dynamics (AA-MD) simulations. To circumvent this problem, a variety of coarse-grained (CG) models have been proposed, ranging from the generic CG models for qualitative properties predictions to more realistic chemically specific CG models for quantitative properties predictions. These CG models have already delivered some success stories in the study of several spatial and temporal evolutions of many processes. Some of these studies were beyond the feasibility of traditional atomistic resolution models due to either the size or the time constraints. This review captures the different types of popular CG approaches that are utilized in the investigation of the microscopic behavior of polymer-nanoparticle hybrid systems. The rationale of this article is to furnish an overview of the popular CG approaches and their applications, to review several important and most recent developments, and to delineate the perspectives on future directions in the field.

Systematic bottom-up molecular coarse-graining via force and torque matching using anisotropic particles

We derive a systematic and general method for parametrizing coarse-grained molecular models consisting of anisotropic particles from fine-grained (e.g. all-atom) models for condensed-phase molecular dynamics simulations. The method, which we call anisotropic force-matching coarse-graining (AFM-CG), is based on rigorous statistical mechanical principles, enforcing consistency between the coarse-grained and fine-grained phase-space distributions to derive equations for the coarse-grained forces, torques, masses, and moments of inertia in terms of properties of a condensed-phase fine-grained system. We verify the accuracy and efficiency of the method by coarse-graining liquid-state systems of two different anisotropic organic molecules, benzene and perylene, and show that the parametrized coarse-grained models more accurately describe properties of these systems than previous anisotropic coarse-grained models parametrized using other methods that do not account for finite-temperature and many-body effects on the condensed-phase coarse-grained interactions. The AFM-CG method will be useful for developing accurate and efficient dynamical simulation models of condensed-phase systems of molecules consisting of large, rigid, anisotropic fragments, such as liquid crystals, organic semiconductors, and nucleic acids.

Automatic multi-objective optimization of coarse-grained lipid force fields using SwarmCG

The development of coarse-grained (CG) molecular models typically requires a time-consuming iterative tuning of parameters in order to have the approximated CG models behave correctly and consistently with, e.g., available higher-resolution simulation data and/or experimental observables. Automatic data-driven approaches are increasingly used to develop accurate models for molecular dynamics simulations. However, the parameters obtained via such automatic methods often make use of specifically designed interaction potentials and are typically poorly transferable to molecular systems or conditions other than those used for training them. Using a multi-objective approach in combination with an automatic optimization engine (SwarmCG), here, we show that it is possible to optimize CG models that are also transferable, obtaining optimized CG force fields (FFs). As a proof of concept, here, we use lipids for which we can avail reference experimental data (area per lipid and bilayer thickness) and reliable atomistic simulations to guide the optimization. Once the resolution of the CG models (mapping) is set as an input, SwarmCG optimizes the parameters of the CG lipid models iteratively and simultaneously against higher-resolution simulations (bottom-up) and experimental data (top-down references). Including different types of lipid bilayers in the training set in a parallel optimization guarantees the transferability of the optimized lipid FF parameters. We demonstrate that SwarmCG can reach satisfactory agreement with experimental data for different resolution CG FFs. We also obtain stimulating insights into the precision-resolution balance of the FFs. The approach is general and can be effectively used to develop new FFs and to improve the existing ones.

Solvent-scaling as an alternative to coarse-graining in adaptive-resolution simulations: The adaptive solvent-scaling (AdSoS) scheme

A new approach termed Adaptive Solvent-Scaling (AdSoS) is introduced for performing simulations of a solute embedded in a fine-grained (FG) solvent region itself surrounded by a coarse-grained (CG) solvent region, with a continuous FG ↔ CG switching of the solvent resolution across a buffer layer. Instead of relying on a distinct CG solvent model, the AdSoS scheme is based on CG models defined by a dimensional scaling of the FG solvent by a factor s, accompanied by an s-dependent modulation of the atomic masses and interaction parameters. The latter changes are designed to achieve an isomorphism between the dynamics of the FG and CG models, and to preserve the dispersive and dielectric solvation properties of the solvent with respect to a solute at FG resolution. This scaling approach offers a number of advantages compared to traditional coarse-graining: (i) the CG parameters are immediately related to those of the FG model (no need to parameterize a distinct CG model); (ii) nearly ideal mixing is expected for CG variants with similar s-values (ideal mixing holding in the limit of identical s-values); (iii) the solvent relaxation timescales should be preserved (no dynamical acceleration typical for coarse-graining); (iv) the graining level N_G (number of FG molecules represented by one CG molecule) can be chosen arbitrarily (in particular, N_G = s³ is not necessarily an integer); and (v) in an adaptive-resolution scheme, this level can be varied continuously as a function of the position (without requiring a bundling mechanism), and this variation occurs at a constant number of particles per molecule (no occurrence of fractional degrees of freedom in the buffer layer). By construction, the AdSoS scheme minimizes the thermodynamic mismatch between the different regions of the adaptive-resolution system, leading to a nearly homogeneous scaled solvent density s³ρ. Residual density artifacts in and at the surface of the boundary layer can easily be corrected by means of a grid-based biasing potential constructed in a preliminary pure-solvent simulation. This article introduces the AdSoS scheme and provides an initial application to pure atomic liquids (no solute) with Lennard-Jones plus Coulomb interactions in a slab geometry.

Multiscale polarizable coarse-graining water models on cluster-level electrostatic dipoles

The development of a coarse-grained (CG) water model is increasingly important in CG studies of biological processes. In this work, we developed a generic CG force field of liquid water on cluster-level electrostatic dipoles. An exponential term is introduced in the non-bonded potential to adjust the well depth. The whole force field is parametrized on the AMOEBA simulation and then refined on the experimental density, dielectric permittivity and isothermal compressibility. The new CG water force field is suitable for the construction of multi-resolution water models and here the N_C = 4/5/10 systems are taken as examples. The results show that the N_C = 4/5/10 models can correctly reproduce the density and relative dielectric permittivity. The models can well predict the pressure-density/density-temperature relationships close to the all-atom or experiment results. However, the new models behave differently from other CG models in several water properties such as the air-water surface tension. Through dipole distributions, two representative polarizable configurations are captured after the N_C = 4/5/10 systems are dynamically equilibrated. Besides, the N_C = 4 model is coupled with the Martini Na⁺/Cl^- models to predict ion-relevant radial distribution functions in comparison to the Martini result. Lastly, CPU tests suggest that the new CG models can enhance simulation efficiency by factors of 20-42, compared to the TIP3P force field. The newly proposed polarizable water force field is practical and transferable and can be flexibly extended to higher coarse-graining of liquid water.

Development of accurate coarse-grained force fields for weakly polar groups by an indirect parameterization strategy

Coarse-grained (CG) molecular dynamics simulations are widely used to predict morphological structures and interpret mechanisms of mesoscopic behavior between the scope of traditional experiments and all-atom simulations. However, most current CG force fields (FFs) are not precise enough, especially for polar molecules or functional groups. A main obstacle in developing accurate CG FFs for polar molecules is the freezing problem met at room temperature. In this work, we introduce an indirect parametrization strategy for weakly polar groups by considering their short-chain homologs to avoid freezing. Here, a polar group containing three to four heavy atoms is mapped into one CG bead that is connected to one alkyl bead composed of three or four carbons. The CG beads interact via 4-parameter nonbonded Morse potentials and harmonic bonded potentials. An efficient meta-multilinear interpolation parameterization algorithm, as recently developed by us, is used to rigorously optimize the force parameters. Satisfactory accuracy is witnessed in terms of the density, heat of vaporization, surface tension, and solvation free energy of the homologs of twelve polar molecules, all deviating from the experiment by less than 5%. The transferability of the current FF is indicated by the predicted density, heat of vaporization, and end-to-end distance distributions of fatty acid methyl esters composed of multiple functional groups parameterized in this work.

Development of coarse-grained force field for alcohols: an efficient meta-multilinear interpolation parameterization algorithm

Coarse-grained (CG) molecular dynamics are powerful tools to access a mesoscopic phenomenon and simultaneously record microscopic details, but currently the CG force fields (FFs) are still limited by low parameterization efficiency and poor accuracy especially for polar molecules. In this work, we developed a Meta-Multilinear Interpolation Parameterization (Meta-MIP) algorithm to optimize the CG FFs for alcohols. This algorithm significantly boosts parameterization efficiency by constructing on-the-fly local databases to cover the global optimal parameterization path. In specific, an alcohol molecule is mapped to a heterologous model composed of an OH bead and a hydrocarbon portion which consists of alkane beads representing two to four carbon atoms. Non-bonded potentials are described by soft Morse functions that have no tail-corrections but can still retain good continuities at truncation distance. Nearly all of the properties in terms of density, heat of vaporization, surface tension, and solvation free energy for alcohols predicted by the current FFs deviate from experimental values by less than 7%. This Meta-MIP algorithm can be readily applied to force field development for a wide variety of molecules or functional groups, in many situations including but not limited to CG FFs.

Swarm-CG: Automatic Parametrization of Bonded Terms in MARTINI-Based Coarse-Grained Models of Simple to Complex Molecules via Fuzzy Self-Tuning Particle Swarm Optimization

We present Swarm-CG, a versatile software for the automatic iterative parametrization of bonded parameters in coarse-grained (CG) models, ideal in combination with popular CG force fields such as MARTINI. By coupling fuzzy self-tuning particle swarm optimization to Boltzmann inversion, Swarm-CG performs accurate bottom-up parametrization of bonded terms in CG models composed of up to 200 pseudo atoms within 4-24 h on standard desktop machines, using default settings. The software benefits from a user-friendly interface and two different usage modes (default and advanced). We particularly expect Swarm-CG to support and facilitate the development of new CG models for the study of complex molecular systems interesting for bio- and nanotechnology. Excellent performances are demonstrated using a benchmark of 9 molecules of diverse nature, structural complexity, and size. Swarm-CG is available with all its dependencies via the Python Package Index (PIP package: swarm-cg). Demonstration data are available at: www.github.com/GMPavanLab/SwarmCG.

A three-point coarse-grained model of five-water cluster with permanent dipoles and quadrupoles

As coarse-grained (CG) studies of large biomolecules increase, developments of reliable CG solvent models become particularly important. In this work, we reduce five water molecules into a three-point CG model with permanent dipole and quadrupole moments. In the CG force field, the modified Morse potential is utilized and an ideal three-water cluster is designed to derive CG-level permanent multipoles. The new CG model is parametrized on the AMOEBA polarizable force field. Various important properties of liquid water are examined to validate the new CG model. Results show that the new CG model can correctly reproduce certain important experimental properties such as density, isothermal compressibility and relative static dielectric permittivity, even better than the existing AA models. Additionally, the CPU tests reveal that the CG model can accelerate molecular dynamics simulations by a factor of 19 compared to the popular AA force field. Compared with the fix-point-charge model widely used in other CG models, the permanent-multipole-based CG model describes more rigid electrostatic attractions. This study also illustrates that the permanent multipole moments contribute a lot to the electrostatic calculations in CG simulation.

Learning Coarse-Grained Potentials for Binary Fluids

For a multiple-fluid system, CG models capable of accurately predicting the interfacial properties as a function of curvature are still lacking. In this work, we propose a new probabilistic machine learning (ML) model for learning CG potentials for binary fluids. The water-hexane mixture is selected as a typical immiscible binary liquid-liquid system. We develop a new CG force field (FF) using the Shinoda-DeVane-Klein (SDK) FF framework and compute parameters in this CG FF using the proposed probabilistic ML method. It is shown that a standard response-surface approach does not provide a unique set of parameters, as it results in a loss function with multiple shallow minima. To address this challenge, we develop a probabilistic ML approach where we compute the probability density function (PDF) of parameters that minimize the loss function. The PDF has a well-defined peak corresponding to a unique set of parameters in the CG FF that reproduces the desired properties of a liquid-liquid interface. We compare the performance of the new CG FF with several existing FFs for the water-hexane mixture, including two atomistic and three CG FFs with respect to modeling the interface structure and thermodynamic properties. It is demonstrated that the new FF significantly improves the CG model prediction of both the interfacial tension and structure for the water-hexane mixture.

Mesoporous Mn-Doped Fe Nanoparticle-Modified Reduced Graphene Oxide for Ethyl Violet Elimination: Modeling and Optimization Using Artificial Intelligence

Mesoporous Mn-doped Fe nanoparticle-modified reduced graphene oxide (Mn-doped Fe/rGO) was prepared through a one-step co-precipitation method, which was then used to eliminate ethyl violet (EV) in wastewater. The prepared Mn-doped Fe/rGO was characterized by X-ray diffraction, X-ray photoelectron spectroscopy, Raman spectroscopy, high-resolution transmission electron microscopy, scanning electron microscopy, energy dispersive spectroscopy, N2-sorption, small angle X-ray diffraction and superconducting quantum interference device. The Brunauer–Emmett–Teller specific surface area of Mn-doped Fe/rGO composites was 104.088 m2/g. The EV elimination by Mn-doped Fe/rGO was modeled and optimized by artificial intelligence (AI) models (i.e., radial basis function network, random forest, artificial neural network genetic algorithm (ANN-GA) and particle swarm optimization). Among these AI models, ANN-GA is considered as the best model for predicting the removal efficiency of EV by Mn-doped Fe/rGO. The evaluation of variables shows that dosage gives the maximum importance to Mn-doped Fe/rGO removal of EV. The experimental data were fitted to kinetics and adsorption isotherm models. The results indicated that the process of EV removal by Mn-doped Fe/rGO obeyed the pseudo-second-order kinetics model and Langmuir isotherm, and the maximum adsorption capacity was 1000.00 mg/g. This study provides a possibility for synthesis of Mn-doped Fe/rGO by co-precipitation as an excellent material for EV removal from the aqueous phase.

Machine-Learning Enabled New Insights into the Coil-to-Globule Transition of Thermosensitive Polymers Using a Coarse-Grained Model

We present a computational framework that integrates coarse-grained (CG) molecular dynamics (MD) simulations and a data-driven machine-learning (ML) method to gain insights into the conformations of polymers in solutions. We employ this framework to study conformational transition of a model thermosensitive polymer, poly( N-isopropylacrylamide) (PNIPAM). Here, we have developed the first of its kind, a temperature-independent CG model of PNIPAM that can accurately predict its experimental lower critical solution temperature (LCST) while retaining the tacticity in the presence of an explicit water model. The CG model was extensively validated by performing CG MD simulations with different initial conformations, varying the radius of gyration of chain, the chain length, and the angle between the adjacent monomers of the initial configuration of PNIPAM (total simulation time = 90 μs). Moreover, for the first time, we utilize the nonmetric multidimensional scaling (NMDS) method, a data-driven ML approach, to gain further insights into the mechanisms and pathways of this coil-to-globule transition by analyzing CG MD simulation trajectories. NMDS analysis provides entirely new insights and shows multiple metastable states of PNIPAM during its coil-to-globule transition above the LCST.

Machine-Learned Coarse-Grained Models

Optimizing force-field (FF) parameters to perform molecular dynamics (MD) simulations is a challenging and time-consuming process. We present a novel FF optimization framework that integrates MD simulations with particle swarm optimization (PSO) algorithm and artificial neural network (ANN). This new ANN-assisted PSO framework was used to develop transferable coarse-grained (CG) models for D₂O and DMF as a proof of concept. The PSO algorithm was used to generate the set of input FF parameters for the MD simulations of the CG models of these solvents, which were optimized to reproduce their experimental properties. Herein, for the first time, a reverse approach was employed for on-the-fly training of the ANN model, where results (solvent properties) obtained from the MD simulations and their corresponding FF parameters were used as inputs and outputs, respectively. The ANN model was then required to predict a set of new FF parameters, which were tested for their ability to predict the desired experimental properties. This new framework can be extended to integrate any optimization algorithm with ANN and MD simulations to accelerate the FF development.