Benchmark assessment of molecular geometries and energies from small molecule force fields

The SAMPL9 host-guest blind challenge: an overview of binding free energy predictive accuracy

We report the results of the SAMPL9 host-guest blind challenge for predicting binding free energies. The challenge focused on macrocycles from pillar[n]-arene and cyclodextrin host families, including WP6, and bCD and HbCD. A variety of methods were used by participants to submit binding free energy predictions. A machine learning approach based on molecular descriptors achieved the highest accuracy (RMSE of 2.04 kcal mol-1) among the ranked methods in the WP6 dataset. Interestingly, predictions for WP6 obtained via docking tended to outperform all methods (RMSE of 1.70 kcal mol-1), most of which are MD based and computationally more expensive. In general, methods applying force fields achieved better correlation with experiments for WP6 opposed to the machine learning and docking models. In the cyclodextrin-phenothiazine challenge, the ATM approach emerged as the top performing method with RMSE less than 1.86 kcal mol-1. Correlation metrics of ranked methods in this dataset were relatively poor compared to WP6. We also highlight several lessons learned to guide future work and help improve studies on the systems discussed. For example, WP6 may be present in other microstates other than its -12 state in the presence of certain guests. Machine learning approaches can be used to fine tune or help train force fields for certain chemistry (i.e. WP6-G4). Certain phenothiazines occupy distinct primary and secondary orientations, some of which were considered individually for accurate binding free energies. The accuracy of predictions from certain methods while starting from a single binding pose/orientation demonstrates the sensitivity of calculated binding free energies to the orientation, and in some cases the likely dominant orientation for the system. Computational and experimental results suggest that guest phenothiazine core traverses both the secondary and primary faces of the cyclodextrin hosts, a bulky cationic side chain will primarily occupy the primary face, and the phenothiazine core substituent resides at the larger secondary face.

Martini on the Rocks: Can a Coarse-Grained Force Field Model Crystals?

Computational chemistry is an important tool in numerous scientific disciplines, including drug discovery and structural biology. Coarse-grained models offer simple representations of molecular systems that enable simulations of large-scale systems. Because there has been an increase in the adoption of such models for simulations of biomolecular systems, critical evaluation is warranted. Here, the stability of the amyloid peptide and organic crystals is evaluated using the Martini 3 coarse-grained force field. The crystals change shape drastically during the simulations. Radial distribution functions show that the distance between backbone beads in β-sheets increases by ∼1 Å, breaking the crystals. The melting points of organic compounds are much too low in the Martini force field. This suggests that Martini 3 lacks the specific interactions needed to accurately simulate peptides or organic crystals without imposing artificial restraints. The problems may be exacerbated by the use of the 12-6 potential, suggesting that a softer potential could improve this model for crystal simulations.

Systematic Comparison of Experimental Crystallographic Geometries and Gas-Phase Computed Conformers for Torsion Preferences

We performed exhaustive torsion sampling on more than 3 million compounds using the GFN2-xTB method and performed a comparison of experimental crystallographic and gas-phase conformers. Many conformer sampling methods derive torsional angle distributions from experimental crystallographic data, limiting the torsion preferences to molecules that must be stable, synthetically accessible, and able to be crystallized. In this work, we evaluate the differences in torsional preferences of experimental crystallographic geometries and gas-phase computed conformers from a broad selection of compounds to determine whether torsional angle distributions obtained from semiempirical methods are suitable priors for conformer sampling. We find that differences in torsion preferences can be mostly attributed to a lack of available experimental crystallographic data with small deviations derived from gas-phase geometry differences. GFN2 demonstrates the ability to provide accurate and reliable torsional preferences that can provide a basis for new methods free from the limitations of experimental data collection. We provide Gaussian-based fits and sampling distributions suitable for torsion sampling and propose an alternative to the widely used "experimental torsion and knowledge distance geometry" (ETKDG) method using quantum torsion-derived distance geometry (QTDG) methods.

Expanded ensemble predictions of absolute binding free energies in the SAMPL9 host-guest challenge

As part of the SAMPL9 community-wide blind host-guest challenge, we implemented an expanded ensemble workflow to predict absolute binding free energies for 13 small molecules against pillar[6]arene. Notable features of our protocol include consideration of a variety of protonation and enantiomeric states for both host and guests, optimization of alchemical intermediates, and analysis of free energy estimates and their uncertainty using large numbers of simulation replicates performed using distributed computing. Our predictions of absolute binding free energies resulted in a mean absolute error of 2.29 kcal mol-1 and an R2 of 0.54. Overall, results show that expanded ensemble calculations using all-atom molecular dynamics simulations are a valuable and efficient computational tool in predicting absolute binding free energies.

ConfSolv: Prediction of Solute Conformer-Free Energies across a Range of Solvents

Predicting Gibbs free energy of solution is key to understanding the solvent effects on thermodynamics and reaction rates for kinetic modeling. Accurately computing solution free energies requires the enumeration and evaluation of relevant solute conformers in solution. However, even after generation of relevant conformers, determining their free energy of solution requires an expensive workflow consisting of several ab initio computational chemistry calculations. To help address this challenge, we generate a large data set of solution free energies for nearly 44,000 solutes with almost 9 million conformers calculated in 41 different solvents using density functional theory and COSMO-RS and quantify the impact of solute conformers on the solution free energy. We then train a message passing neural network to predict the relative solution free energies of a set of solute conformers, enabling the identification of a small subset of thermodynamically relevant conformers. The model offers substantial computational time savings with predictions usually substantially within 1 kcal/mol of the free energy of the solution calculated by using computational chemical methods.

Design, Synthesis, and Biological Evaluation of Novel Quercetin Derivatives as PPAR-γ Partial Agonists by Modulating Epithelial-Mesenchymal Transition in Lung Cancer Metastasis

Epithelial-to-mesenchymal transition (EMT) is responsible for driving metastasis of multiple cancer types including lung cancer. Peroxisome proliferator-activated receptor (PPAR)-γ, a ligand-activated transcription factor, controls expression of variety of genes involved in EMT. Although several synthetic compounds act as potent full agonists for PPAR-γ, their long term application is restricted due to serious adverse effects. Therefore, partial agonists involving reduced and balanced PPAR-γ activity are more effective and valued. A previous study discerned the efficacy of quercetin and its derivatives to attain favorable stabilization with PPAR-γ. Here this work is extended by synthesizing five novel quercetin derivatives (QDs) namely thiosemicarbazone (QUETSC)) and hydrazones (quercetin isonicotinic acid hydrazone (QUEINH), quercetin nicotinic acid hydrazone (QUENH), quercetin 2-furoic hydrazone (QUE2FH), and quercetin salicyl hydrazone (QUESH)) and their effects are analyzed in modulating EMT in lung cancer cell lines via PPAR-γ partial activation. QDs-treated A549 cells diminish cell proliferation strongly at nanomolar concentration compared to NCI-H460 cells. Of the five screened derivatives, QUETSC, QUE2FH, and QUESH exhibit the property of partial activation as compared to the overexpressive level of rosiglitazone. Consistently, these QDs also suppress EMT process by markedly downregulating the levels of mesenchymal markers (Snail, Slug, and zinc finger E-box binding homeobox 1) and concomitant upregulation of epithelial marker (E-cadherin).

A transferable double exponential potential for condensed phase simulations of small molecules

The Lennard-Jones potential is the most widely-used function for the description of non-bonded interactions in transferable force fields for the condensed phase. This is not because it has an optimal functional form, but rather it is a legacy resulting from when computational expense was a major consideration and this potential was particularly convenient numerically. At present, it persists because the effort that would be required to re-write molecular modelling software and train new force fields has, until now, been prohibitive. Here, we present Smirnoff-plugins as a flexible framework to extend the Open Force Field software stack to allow custom force field functional forms. We deploy Smirnoff-plugins with the automated Open Force Field infrastructure to train a transferable, small molecule force field based on the recently-proposed double exponential functional form, on over 1000 experimental condensed phase properties. Extensive testing of the resulting force field shows improvements in transfer free energies, with acceptable conformational energetics, run times and convergence properties compared to state-of-the-art Lennard-Jones based force fields.

Development and Benchmarking of Open Force Field 2.0.0: The Sage Small Molecule Force Field

We introduce the Open Force Field (OpenFF) 2.0.0 small molecule force field for drug-like molecules, code-named Sage, which builds upon our previous iteration, Parsley. OpenFF force fields are based on direct chemical perception, which generalizes easily to highly diverse sets of chemistries based on substructure queries. Like the previous OpenFF iterations, the Sage generation of OpenFF force fields was validated in protein-ligand simulations to be compatible with AMBER biopolymer force fields. In this work, we detail the methodology used to develop this force field, as well as the innovations and improvements introduced since the release of Parsley 1.0.0. One particularly significant feature of Sage is a set of improved Lennard-Jones (LJ) parameters retrained against condensed phase mixture data, the first refit of LJ parameters in the OpenFF small molecule force field line. Sage also includes valence parameters refit to a larger database of quantum chemical calculations than previous versions, as well as improvements in how this fitting is performed. Force field benchmarks show improvements in general metrics of performance against quantum chemistry reference data such as root-mean-square deviations (RMSD) of optimized conformer geometries, torsion fingerprint deviations (TFD), and improved relative conformer energetics (ΔΔE). We present a variety of benchmarks for these metrics against our previous force fields as well as in some cases other small molecule force fields. Sage also demonstrates improved performance in estimating physical properties, including comparison against experimental data from various thermodynamic databases for small molecule properties such as ΔHmix, ρ(x), ΔGsolv, and ΔGtrans. Additionally, we benchmarked against protein-ligand binding free energies (ΔGbind), where Sage yields results statistically similar to previous force fields. All the data is made publicly available along with complete details on how to reproduce the training results at https://github.com/openforcefield/openff-sage.

Does a Machine-Learned Potential Perform Better Than an Optimally Tuned Traditional Force Field? A Case Study on Fluorohydrins

We present a comparative study that evaluates the performance of a machine learning potential (ANI-2x), a conventional force field (GAFF), and an optimally tuned GAFF-like force field in the modeling of a set of 10 γ-fluorohydrins that exhibit a complex interplay between intra- and intermolecular interactions in determining conformer stability. To benchmark the performance of each molecular model, we evaluated their energetic, geometric, and sampling accuracies relative to quantum-mechanical data. This benchmark involved conformational analysis both in the gas phase and chloroform solution. We also assessed the performance of the aforementioned molecular models in estimating nuclear spin-spin coupling constants by comparing their predictions to experimental data available in chloroform. The results and discussion presented in this study demonstrate that ANI-2x tends to predict stronger-than-expected hydrogen bonding and overstabilize global minima and shows problems related to inadequate description of dispersion interactions. Furthermore, while ANI-2x is a viable model for modeling in the gas phase, conventional force fields still play an important role, especially for condensed-phase simulations. Overall, this study highlights the strengths and weaknesses of each model, providing guidelines for the use and future development of force fields and machine learning potentials.

Evidence That Less Can Be More for Transferable Force Fields

Graph-based parameter assignment has been the basis for developing transferable force fields for molecular dynamics simulations for decades. Nevertheless, transferable force fields vary in how specifically terms are defined with respect to the molecular graph and the procedures for generating parametrization data. More-specific force-field terms increase the complexity of the force field, theoretically increasing accuracy but also increasing training data requirements. In contrast, less-specific force fields can be reused across larger regions of chemical space, theoretically reducing accuracy but also reducing the number of parameters and training data requirements. Here, the tradeoffs between force-field specificity and accuracy are quantified by parametrizing three new sets of force fields with varying levels of graph specificity, using a shared procedure for generating training data. These force fields are benchmarked for their ability to reproduce the structural features and liquid properties of 87 organic molecules at 146 distinct state points. The overall accuracy for properties that were directly trained on rapidly saturates as the graph specificity of the force-field increases. From this, we conclude there is at best a marginal benefit of using less transferable and more complex force fields with common sources of quantum-chemically derived training data. When looking at properties unseen during training, there is some evidence that the more-complex force fields even perform slightly worse. These results are rationalized by the fortuitous regularization of force fields based on less-specific and more-transferable atom types. Both the saturation in the accuracy of training properties and the marginally worse performance on off-target properties fundamentally contradict the expectation that bespoke force fields are generally more accurate, given their larger number of parameters, and suggests that increasing force-field complexity should be carefully justified against performance gains and balanced against available training data.

An Imbalance in the Force: The Need for Standardized Benchmarks for Molecular Simulation

Force fields (FFs) for molecular simulation have been under development for more than half a century. As with any predictive model, rigorous testing and comparisons of models critically depends on the availability of standardized data sets and benchmarks. While such benchmarks are rather common in the fields of quantum chemistry, this is not the case for empirical FFs. That is, few benchmarks are reused to evaluate FFs, and development teams rather use their own training and test sets. Here we present an overview of currently available tests and benchmarks for computational chemistry, focusing on organic compounds, including halogens and common ions, as FFs for these are the most common ones. We argue that many of the benchmark data sets from quantum chemistry can in fact be reused for evaluating FFs, but new gas phase data is still needed for compounds containing phosphorus and sulfur in different valence states. In addition, more nonequilibrium interaction energies and forces, as well as molecular properties such as electrostatic potentials around compounds, would be beneficial. For the condensed phases there is a large body of experimental data available, and tools to utilize these data in an automated fashion are under development. If FF developers, as well as researchers in artificial intelligence, would adopt a number of these data sets, it would become easier to compare the relative strengths and weaknesses of different models and to, eventually, restore the balance in the force.

Collaborative Assessment of Molecular Geometries and Energies from the Open Force Field

Force fields form the basis for classical molecular simulations, and their accuracy is crucial for the quality of, for instance, protein-ligand binding simulations in drug discovery. The huge diversity of small-molecule chemistry makes it a challenge to build and parameterize a suitable force field. The Open Force Field Initiative is a combined industry and academic consortium developing a state-of-the-art small-molecule force field. In this report, industry members of the consortium worked together to objectively evaluate the performance of the force fields (referred to here as OpenFF) produced by the initiative on a combined public and proprietary dataset of 19,653 relevant molecules selected from their internal research and compound collections. This evaluation was important because it was completely blind; at most partners, none of the molecules or data were used in force field development or testing prior to this work. We compare the Open Force Field "Sage" version 2.0.0 and "Parsley" version 1.3.0 with GAFF-2.11-AM1BCC, OPLS4, and SMIRNOFF99Frosst. We analyzed force-field-optimized geometries and conformer energies compared to reference quantum mechanical data. We show that OPLS4 performs best, and the latest Open Force Field release shows a clear improvement compared to its predecessors. The performance of established force fields such as GAFF-2.11 was generally worse. While OpenFF researchers were involved in building the benchmarking infrastructure used in this work, benchmarking was done entirely in-house within industrial organizations and the resulting assessment is reported here. This work assesses the force field performance using separate benchmarking steps, external datasets, and involving external research groups. This effort may also be unique in terms of the number of different industrial partners involved, with 10 different companies participating in the benchmark efforts.

Open Force Field BespokeFit: Automating Bespoke Torsion Parametrization at Scale

The development of accurate transferable force fields is key to realizing the full potential of atomistic modeling in the study of biological processes such as protein-ligand binding for drug discovery. State-of-the-art transferable force fields, such as those produced by the Open Force Field Initiative, use modern software engineering and automation techniques to yield accuracy improvements. However, force field torsion parameters, which must account for many stereoelectronic and steric effects, are considered to be less transferable than other force field parameters and are therefore often targets for bespoke parametrization. Here, we present the Open Force Field QCSubmit and BespokeFit software packages that, when combined, facilitate the fitting of torsion parameters to quantum mechanical reference data at scale. We demonstrate the use of QCSubmit for simplifying the process of creating and archiving large numbers of quantum chemical calculations, by generating a dataset of 671 torsion scans for druglike fragments. We use BespokeFit to derive individual torsion parameters for each of these molecules, thereby reducing the root-mean-square error in the potential energy surface from 1.1 kcal/mol, using the original transferable force field, to 0.4 kcal/mol using the bespoke version. Furthermore, we employ the bespoke force fields to compute the relative binding free energies of a congeneric series of inhibitors of the TYK2 protein, and demonstrate further improvements in accuracy, compared to the base force field (MUE reduced from 0.56_0.39^0.77 to 0.42_0.28^0.59 kcal/mol and R² correlation improved from 0.72_0.35^0.87 to 0.93_0.84^0.97).

In Silico Study: Combination of α-Mangostin and Chitosan Conjugated with Trastuzumab against Human Epidermal Growth Factor Receptor 2

Breast cancer is a type of cancer with the highest prevalence worldwide. Almost 10–30% of breast cancer cases are diagnosed as positive for HER2 (human epidermal growth factor receptor 2). The currently available treatment methods still exhibit many shortcomings such as a high incidence of side effects and treatment failure due to resistance. This in silico study aims to simulate α-mangostin and chitosan combination conjugated to trastuzumab formulation against HER2 as an effort to improve breast cancer patient therapy. This molecular docking simulation was done through using PatchDock Server. The materials used including the two-dimensional structure of α-mangostin, chitosan, and sodium tripolyphosphate from the PubChem database; trastuzumab FASTA sequence from the DrugBank database; and HER2 structure obtained from a crystal complex with PDB ID: 1N8Z. The results indicated that the particle of α-mangostin and chitosan combinations interacted mostly with the crystallizable fragment (Fc region) of trastuzumab in the conjugation process. The conjugation of trastuzumab to the particle of a combination of α-mangostin and chitosan resulted in the greatest increase in the binding score of the smallest-sized particles (50 Å) with an increase in the score of 3828 and also gave the most similar mode of interaction with trastuzumab. However, the conjugation of trastuzumab eliminated the similarity of the mode of interaction and increased the value of atomic contact energy. Thus, a cominbation of α-mangostin and chitosan conjugated to a trastuzumab formulation was predicted can increase the effectiveness of breast cancer therapy at a relatively small particle size but with the consequence of decreasing atomic contact energy.

Facile Synthesis of Functionalized Phenoxy Quinolines: Antibacterial Activities against ESBL Producing Escherichia coli and MRSA, Docking Studies, and Structural Features Determination through Computational Approach

The synthesis of new 6-Bromoquinolin-4-ol derivatives (3a–3h) by Chan–Lam coupling utilizing different types of solvents (protic, aprotic, and mixed solvents) and bases was studied in the present manuscript. Furthermore, their potential against ESBL producing Escherichia coli (ESBL E. coli) and methicillin-resistant Staphylococcusaureus (MRSA) were investigated. Commercially available 6-bromoquinolin-4-ol (3a) was reacted with different types of aryl boronic acids along with Cu(OAc)₂ via Chan–Lam coupling methodology utilizing the protic and aprotic and mixed solvents. The molecules (3a–3h) exhibited very good yields with methanol, moderate yields with DMF, and low yields with ethanol solvents, while the mixed solvent CH₃OH/H₂O (8:1) gave more excellent results as compared to the other solvents. The in vitro antiseptic values against ESBL E. coli and MRSA were calculated at five different deliberations (10, 20, 30, 40, 50 mg/well) by agar well diffusion method. The molecule 3e depicted highest antibacterial activity while compounds 3b and 3d showed low antibacterial activity. Additionally, MIC and MBC standards were calculated against the established bacteria by broth dilution method. Furthermore, a molecular docking investigation of the derivatives (3a–3h) were performed. Compound (3e) was highly active and depicted the least binding energy of −5.4. Moreover, to investigate the essential structural and physical properties, the density functional theory (DFT) findings of the synthesized molecules were accomplished by using the basic set PBE0-D3BJ/def2-TZVP/SMD water level of the theory. The synthesized compounds showed an energy gap from 4.93 to 5.07 eV.

Synergistic anti-proliferative activity of JQ1 and GSK2801 in triple-negative breast cancer

BACKGROUND

Triple-negative breast cancer (TNBC) constitutes 10-20% of breast cancers and is challenging to treat due to a lack of effective targeted therapies. Previous studies in TNBC cell lines showed in vitro growth inhibition when JQ1 or GSK2801 were administered alone, and enhanced activity when co-administered. Given their respective mechanisms of actions, we hypothesized the combinatorial effect could be due to the target genes affected. Hence the target genes were characterized for their expression in the TNBC cell lines to prove the combinatorial effect of JQ1 and GSK2801.

METHODS

RNASeq data sets of TNBC cell lines (MDA-MB-231, HCC-1806 and SUM-159) were analyzed to identify the differentially expressed genes in single and combined treatments. The topmost downregulated genes were characterized for their downregulated expression in the TNBC cell lines treated with JQ1 and GSK2801 under different dose concentrations and combinations. The optimal lethal doses were determined by cytotoxicity assays. The inhibitory activity of the drugs was further characterized by molecular modelling studies.

RESULTS

Global expression profiling of TNBC cell lines using RNASeq revealed different expression patterns when JQ1 and GSK2801 were co-administered. Functional enrichment analyses identified several metabolic pathways (i.e., systemic lupus erythematosus, PI3K-Akt, TNF, JAK-STAT, IL-17, MAPK, Rap1 and signaling pathways) enriched with upregulated and downregulated genes when combined JQ1 and GSK2801 treatment was administered. RNASeq identified downregulation of PTPRC, MUC19, RNA5-8S5, KCNB1, RMRP, KISS1 and TAGLN (validated by RT-qPCR) and upregulation of GPR146, SCARA5, HIST2H4A, CDRT4, AQP3, MSH5-SAPCD1, SENP3-EIF4A1, CTAGE4 and RNASEK-C17orf49 when cells received both drugs. In addition to differential gene regulation, molecular modelling predicted binding of JQ1 and GSK2801 with PTPRC, MUC19, KCNB1, TAGLN and KISS1 proteins, adding another mechanism by which JQ1 and GSK2801 could elicit changes in metabolism and proliferation.

CONCLUSION

JQ1-GSK2801 synergistically inhibits proliferation and results in selective gene regulation. Besides suggesting that combinatorial use could be useful therapeutics for the treatment of TNBC, the findings provide a glimpse into potential mechanisms of action for this combination therapy approach.

Automated fitting of transition state force fields for biomolecular simulations

The generation of surrogate potential energy functions (PEF) that are orders of magnitude faster to compute but as accurate as the underlying training data from high-level electronic structure methods is one of the most promising applications of fitting procedures in chemistry. In previous work, we have shown that transition state force fields (TSFFs), fitted to the functional form of MM3* force fields using the quantum guided molecular mechanics (Q2MM) method, provide an accurate description of transition states that can be used for stereoselectivity predictions of small molecule reactions. Here, we demonstrate the applicability of the method for fit TSFFs to the well-established Amber force field, which could be used for molecular dynamics studies of enzyme reaction. As a case study, the fitting of a TSFF to the second hydride transfer in Pseudomonas mevalonii 3-hydroxy-3-methylglutaryl coenzyme A reductase (PmHMGR) is used. The differences and similarities to fitting of small molecule TSFFs are discussed.

A multiscale approach to predict the binding mode of metallo beta-lactamase inhibitors

Antibiotic resistance is a major threat to global public health. β-lactamases, which catalyze breakdown of β-lactam antibiotics, are a principal cause. Metallo β-lactamases (MBLs) represent a particular challenge because they hydrolyze almost all β-lactams and to date no MBL inhibitor has been approved for clinical use. Molecular simulations can aid drug discovery, for example, predicting inhibitor complexes, but empirical molecular mechanics (MM) methods often perform poorly for metalloproteins. Here we present a multiscale approach to model thiol inhibitor binding to IMP-1, a clinically important MBL containing two catalytic zinc ions, and predict the binding mode of a 2-mercaptomethyl thiazolidine (MMTZ) inhibitor. Inhibitors were first docked into the IMP-1 active site, testing different docking programs and scoring functions on multiple crystal structures. Complexes were then subjected to molecular dynamics (MD) simulations and subsequently refined through QM/MM optimization with a density functional theory (DFT) method, B3LYP/6-31G(d), increasing the accuracy of the method with successive steps. This workflow was tested on two IMP-1:MMTZ complexes, for which it reproduced crystallographically observed binding, and applied to predict the binding mode of a third MMTZ inhibitor for which a complex structure was crystallographically intractable. We also tested a 12-6-4 nonbonded interaction model in MD simulations and optimization with a SCC-DFTB QM/MM approach. The results show the limitations of empirical models for treating these systems and indicate the need for higher level calculations, for example, DFT/MM, for reliable structural predictions. This study demonstrates a reliable computational pipeline that can be applied to inhibitor design for MBLs and other zinc-metalloenzyme systems.

Linear Atomic Cluster Expansion Force Fields for Organic Molecules: Beyond RMSE

We demonstrate that fast and accurate linear force fields can be built for molecules using the atomic cluster expansion (ACE) framework. The ACE models parametrize the potential energy surface in terms of body-ordered symmetric polynomials making the functional form reminiscent of traditional molecular mechanics force fields. We show that the four- or five-body ACE force fields improve on the accuracy of the empirical force fields by up to a factor of 10, reaching the accuracy typical of recently proposed machine-learning-based approaches. We not only show state of the art accuracy and speed on the widely used MD17 and ISO17 benchmark data sets, but we also go beyond RMSE by comparing a number of ML and empirical force fields to ACE on more important tasks such as normal-mode prediction, high-temperature molecular dynamics, dihedral torsional profile prediction, and even bond breaking. We also demonstrate the smoothness, transferability, and extrapolation capabilities of ACE on a new challenging benchmark data set comprised of a potential energy surface of a flexible druglike molecule.

Generation of Quantum Configurational Ensembles Using Approximate Potentials

Conformational analysis is of paramount importance in drug design: it is crucial to determine pharmacological properties, understand molecular recognition processes, and characterize the conformations of ligands when unbound. Molecular Mechanics (MM) simulation methods, such as Monte Carlo (MC) and molecular dynamics (MD), are usually employed to generate ensembles of structures due to their ability to extensively sample the conformational space of molecules. The accuracy of these MM-based schemes strongly depends on the functional form of the force field (FF) and its parametrization, components that often hinder their performance. High-level methods, such as ab initio MD, provide reliable structural information but are still too computationally expensive to allow for extensive sampling. Therefore, to overcome these limitations, we present a multilevel MC method that is capable of generating quantum configurational ensembles while keeping the computational cost at a minimum. We show that FF reparametrization is an efficient route to generate FFs that reproduce QM results more closely, which, in turn, can be used as low-cost models to achieve the gold standard QM accuracy. We demonstrate that the MC acceptance rate is strongly correlated with various phase space overlap measurements and that it constitutes a robust metric to evaluate the similarity between the MM and QM levels of theory. As a more advanced application, we present a self-parametrizing version of the algorithm, which combines sampling and FF parametrization in one scheme, and apply the methodology to generate the QM/MM distribution of a ligand in aqueous solution.

Development and Benchmarking of Open Force Field v1.0.0-the Parsley Small-Molecule Force Field

We present a methodology for defining and optimizing a general force field for classical molecular simulations, and we describe its use to derive the Open Force Field 1.0.0 small-molecule force field, codenamed Parsley. Rather than using traditional atom typing, our approach is built on the SMIRKS-native Open Force Field (SMIRNOFF) parameter assignment formalism, which handles increases in the diversity and specificity of the force field definition without needlessly increasing the complexity of the specification. Parameters are optimized with the ForceBalance tool, based on reference quantum chemical data that include torsion potential energy profiles, optimized gas-phase structures, and vibrational frequencies. These quantum reference data are computed and are maintained with QCArchive, an open-source and freely available distributed computing and database software ecosystem. In this initial application of the method, we present essentially a full optimization of all valence parameters and report tests of the resulting force field against compounds and data types outside the training set. These tests show improvements in optimized geometries and conformational energetics and demonstrate that Parsley's accuracy for liquid properties is similar to that of other general force fields, as is accuracy on binding free energies. We find that this initial Parsley force field affords accuracy similar to that of other general force fields when used to calculate relative binding free energies spanning 199 protein-ligand systems. Additionally, the resulting infrastructure allows us to rapidly optimize an entirely new force field with minimal human intervention.

Impact of the Characteristics of Quantum Chemical Databases on Machine Learning Prediction of Tautomerization Energies

An essential aspect for adequate predictions of chemical properties by machine learning models is the database used for training them. However, studies that analyze how the content and structure of the databases used for training impact the prediction quality are scarce. In this work, we analyze and quantify the relationships learned by a machine learning model (Neural Network) trained on five different reference databases (QM9, PC9, ANI-1E, ANI-1, and ANI-1x) to predict tautomerization energies from molecules in Tautobase. For this, characteristics such as the number of heavy atoms in a molecule, number of atoms of a given element, bond composition, or initial geometry on the quality of the predictions are considered. The results indicate that training on a chemically diverse database is crucial for obtaining good results and also that conformational sampling can partly compensate for limited coverage of chemical diversity. The overall best-performing reference database (ANI-1x) performs on average by 1 kcal/mol better than PC9, which, however, contains about 2 orders of magnitude fewer reference structures. On the other hand, PC9 is chemically more diverse by a factor of ∼5 as quantified by the number of atom-in-molecule-based fragments (amons) it contains compared with the ANI family of databases. A quantitative measure for deficiencies is the Kullback-Leibler divergence between reference and target distributions. It is explicitly demonstrated that when certain types of bonds need to be covered in the target database (Tautobase) but are undersampled in the reference databases, the resulting predictions are poor. Examples of this include the poor performance of all databases analyzed to predict C(sp²)-C(sp²) double bonds close to heteroatoms and azoles containing N-N and N-O bonds. Analysis of the results with a Tree MAP algorithm provides deeper understanding of specific deficiencies in predicting tautomerization energies by the reference datasets due to inadequate coverage of chemical space. Capitalizing on this information can be used to either improve existing databases or generate new databases of sufficient diversity for a range of machine learning (ML) applications in chemistry.

Improving small molecule force fields by identifying and characterizing small molecules with inconsistent parameters

Many molecular simulation methods use force fields to help model and simulate molecules and their behavior in various environments. Force fields are sets of functions and parameters used to calculate the potential energy of a chemical system as a function of the atomic coordinates. Despite the widespread use of force fields, their inadequacies are often thought to contribute to systematic errors in molecular simulations. Furthermore, different force fields tend to give varying results on the same systems with the same simulation settings. Here, we present a pipeline for comparing the geometries of small molecule conformers. We aimed to identify molecules or chemistries that are particularly informative for future force field development because they display inconsistencies between force fields. We applied our pipeline to a subset of the eMolecules database, and highlighted molecules that appear to be parameterized inconsistently across different force fields. We then identified over-represented functional groups in these molecule sets. The molecules and moieties identified by this pipeline may be particularly helpful for future force field parameterization.

Evaluating the Performance of Water Models with Host-Guest Force Fields in Binding Enthalpy Calculations for Cucurbit[7]uril-Guest Systems

Computational prediction of thermodynamic components with computational methods has become increasingly routine in computer-aided drug design. Although there has been significant recent effort and improvements in the calculation of free energy, the prediction of enthalpy (and entropy) remains underexplored. Furthermore, there has been relatively little work reported so far that attempts to comparatively assess how well different force fields and water models perform in conjunction with each other. Here, we report a comprehensive assessment of force fields and water models using host-guest systems that mimic many features of protein-ligand systems. These systems are computationally inexpensive, possibly because of their small size compared to protein-ligand systems. We present absolute enthalpy calculations using the multibox approach on a set of 25 cucurbit[7]uril-guest pairs. Eight water models were considered (TIP3P, TIP4P, TIP4P-Ew, SPC, SPC/E, OPC, TIP5P, Bind3P), along with five force fields commonly used in the literature (GAFFv1, GAFFv2, CGenFF, Parsley, and SwissParam). We observe that host-guest binding enthalpies are strongly sensitive to the selection of force field and water model. In terms of water models, we find that TIP3P and its derivative Bind3P are the best performing models for this particular host-guest system. The performance is generally better for aliphatic compounds than for aromatic ones, suggesting that aromaticity remains a difficult property to include accurately in these simple force fields.