Toward Achieving Efficient and Accurate Ligand-Protein Unbinding with Deep Learning and Molecular Dynamics through RAVE

Exploring Kinase Asp-Phe-Gly (DFG) Loop Conformational Stability with AlphaFold2-RAVE

Kinases compose one of the largest fractions of the human proteome, and their misfunction is implicated in many diseases, in particular, cancers. The ubiquitousness and structural similarities of kinases make specific and effective drug design difficult. In particular, conformational variability due to the evolutionarily conserved Asp-Phe-Gly (DFG) motif adopting in and out conformations and the relative stabilities thereof are key in structure-based drug design for ATP competitive drugs. These relative conformational stabilities are extremely sensitive to small changes in sequence and provide an important problem for sampling method development. Since the invention of AlphaFold2, the world of structure-based drug design has noticeably changed. In spite of it being limited to crystal-like structure prediction, several methods have also leveraged its underlying architecture to improve dynamics and enhanced sampling of conformational ensembles, including AlphaFold2-RAVE. Here, we extend AlphaFold2-RAVE and apply it to a set of kinases: the wild type DDR1 sequence and three mutants with single point mutations that are known to behave drastically differently. We show that AlphaFold2-RAVE is able to efficiently recover the changes in relative stability using transferable learned order parameters and potentials, thereby supplementing AlphaFold2 as a tool for exploration of Boltzmann-weighted protein conformations (Meller, A.; Bhakat, S.; Solieva, S.; Bowman, G. R. Accelerating Cryptic Pocket Discovery Using AlphaFold. J. Chem. Theory Comput. 2023, 19, 4355-4363).

Enhanced Sampling for Conformational Changes and Molecular Mechanisms of Human NTHL1

The functionalities of proteins rely on protein conformational changes during many processes. Identification of the protein conformations and capturing transitions among different conformations are important but extremely challenging in both experiments and simulations. In this work, we develop a machine learning based approach to identify a reaction coordinate that accelerates the exploration of protein conformational changes in molecular simulations. We implement our approach to study the conformational changes of human NTHL1 during DNA repair. Our results identified three distinct conformations: open (stable), closed (unstable), and bundle (stable). The existence of the bundle conformation can rationalize recent experimental observations. Comparison with an NTHL1 mutant demonstrates that a closely packed cluster of positively charged residues in the linker could be a factor to search when screening for genetic abnormalities. Results will lead to a better modulation of the DNA repair pathway to protect against carcinogenesis.

Excited States of apo-Guanidine-III Riboswitch Contribute to Guanidinium Binding through Both Conformational and Induced-Fit Mechanisms

Riboswitches are mRNA segments that regulate gene expression through conformational changes driven by their cognate ligand binding. The ykkC motif forms a riboswitch class that selectively senses a guanidinium ion (Gdm+) and regulates the downstream expression of proteins which aid in the efflux of excess Gdm+ from the cells. The aptamer domain (AD) of the guanidine-III riboswitch forms an H-type pseudoknot with a triple helical domain that binds a Gdm+. We studied the binding of Gdm+ to the AD of the guanidine (ykkC)-III riboswitch using computer simulations to probe the specificity of the riboswitch to Gdm+ binding. We show that Gdm+ binding is a fast process occurring on the nanosecond time scale, with minimal conformational changes to the AD. Using machine learning and Markov-state models, we identified the excited conformational states of the AD, which have a high Gdm+ binding propensity, making the Gdm+ binding landscape complex exhibiting both conformational selection and induced-fit mechanisms. The proposed apo-AD excited states and their role in the ligand-sensing mechanism are amenable to experimental verification. Further, targeting these excited-state conformations in discovering new antibiotics can be explored.

Deep-Learning-Assisted Enhanced Sampling for Exploring Molecular Conformational Changes

We present a novel strategy to explore conformational changes and identify stable states of molecular objects, eliminating the need for a priori knowledge. The approach applies a deep learning method to extract information about the movement modes of the molecular object from a short, high-dimensional, and parameter-free preliminary enhanced-sampling simulation. The gathered information is described by a small set of deep-learning-based collective variables (dCVs), which steer the production-enhanced-sampling simulation. Considering the challenge of adequately exploring the configurational space using the low-dimensional, suboptimal dCVs, we incorporate a method designed for ergodic sampling, namely, Gaussian-accelerated molecular dynamics (MD), into the framework of CV-based enhanced sampling. MD simulations on both toy models and nontrivial examples demonstrate the remarkable computational efficiency of the strategy in capturing the conformational changes of molecular objects without a priori knowledge. Specifically, we achieved the blind folding of two fast folders, chignolin and villin, within a time scale of hundreds of nanoseconds and successfully reconstructed the free-energy landscapes that characterize their reversible folding. All in all, the presented strategy holds significant promise for investigating conformational changes in macromolecules, and it is anticipated to find extensive applications in the fields of chemistry and biology.

Prediction of the Peptide-TIM3 Binding Site in Inhibiting TIM3-Galectin 9 Binding Pathways

T-cell immunoglobulin and mucin domain-containing protein-3 (TIM3) is an important receptor protein that modulates the immune system. The binding of TIM3 with Galectin 9 (GAL9) triggers immune system suppression, but the TIM3-GAL9 binding can be inhibited by binding of the peptide P26 to TIM3. A fast and accurate prediction of the P26-TIM3 binding site is crucial and a prerequisite for the investigation of P26-TIM3 interactions and TIM3-GAL9 binding pathways. Here, we present a machine learning approach, which considers protein conformational changes, to quickly identify the ligand-binding site on TIM3. Our results show that the P26 binding site is located near the C″-D loop of TIM3. Further simulations show that the binding pose is stabilized by a variety of electrostatic and hydrophobic interactions. Binding of P26 can alter the conformations of nearby glycan side chains on TIM3, providing possible mechanisms of how P26 inhibits TIM3-GAL9 binding pathways. The insights from this work will facilitate the identification of other peptides or antibodies that may also inhibit the TIM3-GAL9 pathways and eventually lead to improved attempts in the modulation of the TIM3-GAL9 immunosuppression pathways. The strategies and machine learning method can be generalized to study ligand-receptor binding when the conformational changes during the binding are important.

15 Years of molecular simulation of drug-binding kinetics

INTRODUCTION

Drug-binding kinetics has been increasingly recognized as an important factor to be considered in drug discovery. Long residence time could prolong the action of some drugs while produce toxicity on others. Early evaluation of the binding kinetics of drug candidates could reduce attrition rate late in the drug discovery process. Computational prediction of drug-binding kinetics is useful as compounds can be evaluated even before they are made. However, simulation of drug-binding kinetics is a challenging problem because of the long-time scale involved. Nevertheless, significant progress has been made.

AREAS COVERED

This review illustrates the rapid evolution of qualitative to quantitative molecular dynamics-based methods that have been developed over the last 15 years.

EXPERT OPINION

The development of new methods based on molecular dynamics simulations now enables computation of absolute association/dissociation rate constants. Cheaper methods capable of identifying candidates with fast or slow binding kinetics, or rank-ordering rate constants are also available. Together, these methods have generated useful insights into the molecular mechanisms of drug-binding kinetics, and the design of drug candidates with therapeutically favorable kinetics. Although predicting absolute rate constants is still expensive and challenging, rapid improvement is expected in the coming years with the continuing refinement of current technologies, development of new methodologies, and the utilization of machine learning.

A deep encoder-decoder framework for identifying distinct ligand binding pathways

The pathway(s) that a ligand would adopt en route to its trajectory to the native pocket of the receptor protein act as a key determinant of its biological activity. While Molecular Dynamics (MD) simulations have emerged as the method of choice for modeling protein-ligand binding events, the high dimensional nature of the MD-derived trajectories often remains a barrier in the statistical elucidation of distinct ligand binding pathways due to the stochasticity inherent in the ligand's fluctuation in the solution and around the receptor. Here, we demonstrate that an autoencoder based deep neural network, trained using an objective input feature of a large matrix of residue-ligand distances, can efficiently produce an optimal low-dimensional latent space that stores necessary information on the ligand-binding event. In particular, for a system of L99A mutant of T4 lysozyme interacting with its native ligand, benzene, this deep encoder-decoder framework automatically identifies multiple distinct recognition pathways, without requiring user intervention. The intermediates involve the spatially discrete location of the ligand in different helices of the protein before its eventual recognition of native pose. The compressed subspace derived from the autoencoder provides a quantitatively accurate measure of the free energy and kinetics of ligand binding to the native pocket. The investigation also recommends that while a linear dimensional reduction technique, such as time-structured independent component analysis, can do a decent job of state-space decomposition in cases where the intermediates are long-lived, autoencoder is the method of choice in systems where transient, low-populated intermediates can lead to multiple ligand-binding pathways.

Predicting Protein-Ligand Binding and Unbinding Kinetics with Biased MD Simulations and Coarse-Graining of Dynamics: Current State and Challenges

The prediction of drug-target binding and unbinding kinetics that occur on time scales between milliseconds and several hours is a prime challenge for biased molecular dynamics simulation approaches. This Perspective gives a concise summary of the theory and the current state-of-the-art of such predictions via biased simulations, of insights into the molecular mechanisms defining binding and unbinding kinetics as well as of the extraordinary challenges predictions of ligand kinetics pose in comparison to binding free energy predictions.

Predicting Biomolecular Binding Kinetics: A Review

Biomolecular binding kinetics including the association (k_on) and dissociation (k_off) rates are critical parameters for therapeutic design of small-molecule drugs, peptides, and antibodies. Notably, the drug molecule residence time or dissociation rate has been shown to correlate with their efficacies better than binding affinities. A wide range of modeling approaches including quantitative structure-kinetic relationship models, Molecular Dynamics simulations, enhanced sampling, and Machine Learning has been developed to explore biomolecular binding and dissociation mechanisms and predict binding kinetic rates. Here, we review recent advances in computational modeling of biomolecular binding kinetics, with an outlook for future improvements.

Ligand Gaussian Accelerated Molecular Dynamics 2 (LiGaMD2): Improved Calculations of Ligand Binding Thermodynamics and Kinetics with Closed Protein Pocket

Ligand binding thermodynamics and kinetics are critical parameters for drug design. However, it has proven challenging to efficiently predict ligand binding thermodynamics and kinetics from molecular simulations due to limited simulation timescales. Protein dynamics, especially in the ligand binding pocket, often plays an important role in ligand binding. Based on our previously developed Ligand Gaussian accelerated molecular dynamics (LiGaMD), here we present LiGaMD2 in which a selective boost potential was applied to both the ligand and protein residues in the binding pocket to improve sampling of ligand binding and dissociation. To validate the performance of LiGaMD2, the T4 lysozyme (T4L) mutants with open and closed pockets bound by different ligands were chosen as model systems. LiGaMD2 could efficiently capture repetitive ligand dissociation and binding within microsecond simulations of all T4L systems. The obtained ligand binding kinetic rates and free energies agreed well with available experimental values and previous modeling results. Therefore, LiGaMD2 provides an improved approach to sample opening of closed protein pockets for ligand dissociation and binding, thereby allowing for efficient calculations of ligand binding thermodynamics and kinetics.

Transferring chemical and energetic knowledge between molecular systems with machine learning

Predicting structural and energetic properties of a molecular system is one of the fundamental tasks in molecular simulations, and it has applications in chemistry, biology, and medicine. In the past decade, the advent of machine learning algorithms had an impact on molecular simulations for various tasks, including property prediction of atomistic systems. In this paper, we propose a novel methodology for transferring knowledge obtained from simple molecular systems to a more complex one, endowed with a significantly larger number of atoms and degrees of freedom. In particular, we focus on the classification of high and low free-energy conformations. Our approach relies on utilizing (i) a novel hypergraph representation of molecules, encoding all relevant information for characterizing multi-atom interactions for a given conformation, and (ii) novel message passing and pooling layers for processing and making free-energy predictions on such hypergraph-structured data. Despite the complexity of the problem, our results show a remarkable Area Under the Curve of 0.92 for transfer learning from tri-alanine to the deca-alanine system. Moreover, we show that the same transfer learning approach can also be used in an unsupervised way to group chemically related secondary structures of deca-alanine in clusters having similar free-energy values. Our study represents a proof of concept that reliable transfer learning models for molecular systems can be designed, paving the way to unexplored routes in prediction of structural and energetic properties of biologically relevant systems.

LINES: Log-Probability Estimation via Invertible Neural Networks for Enhanced Sampling

It is very challenging to sample a molecular process with large activation energies using molecular dynamics simulations. Current enhanced sampling methodologies, such as umbrella sampling and metadynamics, rely on the identification of appropriate reaction coordinates for a system. In this paper, we developed a method for log-probability estimation via invertible neural networks for enhanced sampling (LINES). This iterative scheme utilizes a normalizing flow machine learning model to learn the underlying free energy surface (FES) of a system as a function of molecular coordinates and then applies a gradient-based optimization method to the learned normalizing flow to identify reaction coordinates. A biasing potential is then evaluated over a tabulated grid of the reaction coordinate values, which can be applied to the next round of simulations for enhanced sampling, resulting in more efficient sampling. We tested the accuracy and efficiency of the LINES method in sampling the FES using the alanine dipeptide system. We also demonstrated the effectiveness of identification of reaction coordinates through simulation of cyclobutanol unbinding from β-cyclodextrin and the folding/unfolding of CLN025─a variant of the peptide Chignolin. The LINES method can be extended to the study of large-scale protein systems with complex nonlinear reaction pathways.

Collective variable discovery in the age of machine learning: reality, hype and everything in between

Understanding the kinetics and thermodynamics profile of biomolecules is necessary to understand their functional roles which has a major impact in mechanism driven drug discovery. Molecular dynamics simulation has been routinely used to understand conformational dynamics and molecular recognition in biomolecules. Statistical analysis of high-dimensional spatiotemporal data generated from molecular dynamics simulation requires identification of a few low-dimensional variables which can describe the essential dynamics of a system without significant loss of information. In physical chemistry, these low-dimensional variables are often called collective variables. Collective variables are used to generate reduced representations of free energy surfaces and calculate transition probabilities between different metastable basins. However the choice of collective variables is not trivial for complex systems. Collective variables range from geometric criteria such as distances and dihedral angles to abstract ones such as weighted linear combinations of multiple geometric variables. The advent of machine learning algorithms led to increasing use of abstract collective variables to represent biomolecular dynamics. In this review, I will highlight several nuances of commonly used collective variables ranging from geometric to abstract ones. Further, I will put forward some cases where machine learning based collective variables were used to describe simple systems which in principle could have been described by geometric ones. Finally, I will put forward my thoughts on artificial general intelligence and how it can be used to discover and predict collective variables from spatiotemporal data generated by molecular dynamics simulations.

Systematic Control of Collective Variables Learned from Variational Autoencoders

Variational autoencoders (VAEs) are rapidly gaining popularity within molecular simulation for discovering low-dimensional, or latent, representations, which are critical for both analyzing and accelerating simulations. However, it remains unclear how the information a VAE learns is connected to its probabilistic structure, and, in turn, its loss function. Previous studies have focused on feature engineering, \emph{ad hoc} modifications to loss functions, or adjustment of the prior to enforce desirable latent space properties. By applying effectively arbitrarily flexible priors via normalizing flows, we focus instead on how adjusting the structure of the decoding model impacts the learned latent coordinate. We systematically adjust the power and flexibility of the decoding distribution, observing that this has a significant impact on the structure of the latent space as measured by a suite of metrics developed in this work. By also varying weights on separate terms within each VAE loss function, we show that the level of detail encoded can be further tuned. This provides practical guidance for utilizing VAEs to extract varying resolutions of low-dimensional information from molecular dynamics and Monte Carlo simulations.

Classical and Machine Learning Methods for Protein - Ligand Binding Free Energy Estimation

Binding free energy estimation of drug candidates to their biomolecular target is one of the best quantitative estimators in computer-aided drug discovery. Accurate binding free energy estimation is still a challengeable task even after decades of research, along with the complexity of the algorithm, time-consuming procedures, and reproducibility issues. In this review, we have discussed the advantages and disadvantages of diverse free energy methods like Thermodynamic Integration (TI), Bennett's Acceptance Ratio (BAR), Free Energy Perturbation (FEP), alchemical methods. Moreover, we discussed the possible application of the machine learning method in protein-ligand binding free energy estimation.

Combined Free-Energy Calculation and Machine Learning Methods for Understanding Ligand Unbinding Kinetics

The determination of drug residence times, which define the time an inhibitor is in complex with its target, is a fundamental part of the drug discovery process. Synthesis and experimental measurements of kinetic rate constants are, however, expensive and time consuming. In this work, we aimed to obtain drug residence times computationally. Furthermore, we propose a novel algorithm to identify molecular design objectives based on ligand unbinding kinetics. We designed an enhanced sampling technique to accurately predict the free-energy profiles of the ligand unbinding process, focusing on the free-energy barrier for unbinding. Our method first identifies unbinding paths determining a corresponding set of internal coordinates (ICs) that form contacts between the protein and the ligand; it then iteratively updates these interactions during a series of biased molecular dynamics (MD) simulations to reveal the ICs that are important for the whole of the unbinding process. Subsequently, we performed finite-temperature string simulations to obtain the free-energy barrier for unbinding using the set of ICs as a complex reaction coordinate. Importantly, we also aimed to enable the further design of drugs focusing on improved residence times. To this end, we developed a supervised machine learning (ML) approach with inputs from unbiased “downhill” trajectories initiated near the transition state (TS) ensemble of the string unbinding path. We demonstrate that our ML method can identify key ligand–protein interactions driving the system through the TS. Some of the most important drugs for cancer treatment are kinase inhibitors. One of these kinase targets is cyclin-dependent kinase 2 (CDK2), an appealing target for anticancer drug development. Here, we tested our method using two different CDK2 inhibitors for the potential further development of these compounds. We compared the free-energy barriers obtained from our calculations with those observed in available experimental data. We highlighted important interactions at the distal ends of the ligands that can be targeted for improved residence times. Our method provides a new tool to determine unbinding rates and to identify key structural features of the inhibitors that can be used as starting points for novel design strategies in drug discovery.

Nonlinear Reaction Coordinate of an Enzyme Catalyzed Proton Transfer Reaction

We present an in-depth study on the theoretical calculation of an optimum reaction coordinate as a linear or nonlinear combination of important collective variables (CVs) sampled from an ensemble of reactive transition paths for an intramolecular proton transfer reaction catalyzed by the enzyme human carbonic anhydrase (HCA) II. The linear models are optimized by likelihood maximization for a given number of CVs. The nonlinear models are based on an artificial neural network with the same number of CVs and optimized by minimizing the root-mean-square error in comparison to a training set of committor estimators generated for the given transition. The nonlinear reaction coordinate thus obtained yields the free energy of activation and rate constant as 9.46 kcal mol^-1 and 1.25 × 10⁶ s^-1, respectively. These estimates are found to be in quantitative agreement with the known experimental results. We have also used an extended autoencoder model to show that a similar analysis can be carried out using a single CV only. The resultant free energies and kinetics of the reaction slightly overestimate the experimental data. The implications of these results are discussed using a detailed microkinetic scheme of the proton transfer reaction catalyzed by HCA II.

MLCV: Bridging Machine-Learning-Based Dimensionality Reduction and Free-Energy Calculation

Importance-sampling algorithms leaning on the definition of a model reaction coordinate (RC) are widely employed to probe processes relevant to chemistry and biology alike, spanning time scales not amenable to common, brute-force molecular dynamics (MD) simulations. In practice, the model RC often consists of a handful of collective variables (CVs) chosen on the basis of chemical intuition. However, constructing manually a low-dimensional RC model to describe an intricate geometrical transformation for the purpose of free-energy calculations and analyses remains a daunting challenge due to the inherent complexity of the conformational transitions at play. To solve this issue, remarkable progress has been made in employing machine-learning techniques, such as autoencoders, to extract the low-dimensional RC model from a large set of CVs. Implementation of the differentiable, nonlinear machine-learned CVs in common MD engines to perform free-energy calculations is, however, particularly cumbersome. To address this issue, we present here a user-friendly tool (called MLCV) that facilitates the use of machine-learned CVs in importance-sampling simulations through the popular Colvars module. Our approach is critically probed with three case examples consisting of small peptides, showcasing that through hard-coded neural network in Colvars, deep-learning and enhanced-sampling can be effectively bridged with MD simulations. The MLCV code is versatile, applicable to all the CVs available in Colvars, and can be connected to any kind of dense neural networks. We believe that MLCV provides an effective, powerful, and user-friendly platform accessible to experts and nonexperts alike for machine-learning (ML)-guided CV discovery and enhanced-sampling simulations to unveil the molecular mechanisms underlying complex biochemical processes.

The prediction of protein-ligand unbinding for modern drug discovery

INTRODUCTION

Drug-target thermodynamic and kinetic information have perennially important roles in drug design. The prediction of protein-ligand unbinding, which can provide important kinetic information, in experiments continues to face great challenges. Uncovering protein-ligand unbinding through molecular dynamics simulations has become efficient and inexpensive with the progress and enhancement of computing power and sampling methods.

AREAS COVERED

In this review, various sampling methods for protein-ligand unbinding and their basic principles are firstly briefly introduced. Then, their applications in predicting aspects of protein-ligand unbinding, including unbinding pathways, dissociation rate constants, residence time and binding affinity, are discussed.

EXPERT OPINION

Although various sampling methods have been successfully applied in numerous systems, they still have shortcomings and deficiencies. Most enhanced sampling methods require researchers to possess a wealth of prior knowledge of collective variables or reaction coordinates. In addition, most systems studied at present are relatively simple, and the study of complex systems in real drug research remains greatly challenging. Through the combination of machine learning and enhanced sampling methods, prediction accuracy can be further improved, and some problems encountered in complex systems also may be solved.

Energy penalties enhance flexible receptor docking in a model cavity

Protein flexibility remains a major challenge in library docking because of difficulties in sampling conformational ensembles with accurate probabilities. Here, we use the model cavity site of T4 lysozyme L99A to test flexible receptor docking with energy penalties from molecular dynamics (MD) simulations. Crystallography with larger and smaller ligands indicates that this cavity can adopt three major conformations: open, intermediate, and closed. Since smaller ligands typically bind better to the cavity site, we anticipate an energy penalty for the cavity opening. To estimate its magnitude, we calculate conformational preferences from MD simulations. We find that including a penalty term is essential for retrospective ligand enrichment; otherwise, high-energy states dominate the docking. We then prospectively docked a library of over 900,000 compounds for new molecules binding to each conformational state. Absent a penalty term, the open conformation dominated the docking results; inclusion of this term led to a balanced sampling of ligands against each state. High ranked molecules were experimentally tested by Tm upshift and X-ray crystallography. From 33 selected molecules, we identified 18 ligands and determined 13 crystal structures. Most interesting were those bound to the open cavity, where the buried site opens to bulk solvent. Here, highly unusual ligands for this cavity had been predicted, including large ligands with polar tails; these were confirmed both by binding and by crystallography. In docking, incorporating protein flexibility with thermodynamic weightings may thus access new ligand chemotypes. The MD approach to accessing and, crucially, weighting such alternative states may find general applicability.

AI-driven multiscale simulations illuminate mechanisms of SARS-CoV-2 spike dynamics

We develop a generalizable AI-driven workflow that leverages heterogeneous HPC resources to explore the time-dependent dynamics of molecular systems. We use this workflow to investigate the mechanisms of infectivity of the SARS-CoV-2 spike protein, the main viral infection machinery. Our workflow enables more efficient investigation of spike dynamics in a variety of complex environments, including within a complete SARS-CoV-2 viral envelope simulation, which contains 305 million atoms and shows strong scaling on ORNL Summit using NAMD. We present several novel scientific discoveries, including the elucidation of the spike's full glycan shield, the role of spike glycans in modulating the infectivity of the virus, and the characterization of the flexible interactions between the spike and the human ACE2 receptor. We also demonstrate how AI can accelerate conformational sampling across different systems and pave the way for the future application of such methods to additional studies in SARS-CoV-2 and other molecular systems.

Ligand unbinding mechanisms and kinetics for T4 lysozyme mutants from τRAMD simulations

The protein-ligand residence time, τ, influences molecular function in biological networks and has been recognized as an important determinant of drug efficacy. To predict τ, computational methods must overcome the problem that τ often exceeds the timescales accessible to conventional molecular dynamics (MD) simulation. Here, we apply the τ-Random Acceleration Molecular Dynamics (τRAMD) method to a set of kinetically characterized complexes of T4 lysozyme mutants with small, engineered binding cavities. τRAMD yields relative ligand dissociation rates in good accordance with experiments across this diverse set of complexes that differ with regard to measurement temperature, ligand identity, protein mutation and binding cavity. τRAMD thereby allows a comprehensive characterization of the ligand egress routes and determinants of τ. Although ligand dissociation by multiple egress routes is observed, we find that egress via the predominant route determines the value of τ. We also find that the presence of a greater number of metastable states along egress pathways leads to slower protein-ligand dissociation. These physical insights could be exploited in the rational optimization of the kinetic properties of drug candidates.

•

Relative residence times are computed for T4 lysozyme mutant-ligand complexes.

•

τ-Random Acceleration Molecular Dynamics provide efficient sampling of unbinding.

•

Computed dissociation rates show good agreement with all available measured values.

•

Ligand egress via the predominant route determines the value of the residence time.

•

The presence of metastable states along egress pathways slows down dissociation.

Artificial intelligence techniques for integrative structural biology of intrinsically disordered proteins

We outline recent developments in artificial intelligence (AI) and machine learning (ML) techniques for integrative structural biology of intrinsically disordered proteins (IDP) ensembles. IDPs challenge the traditional protein structure-function paradigm by adapting their conformations in response to specific binding partners leading them to mediate diverse, and often complex cellular functions such as biological signaling, self-organization and compartmentalization. Obtaining mechanistic insights into their function can therefore be challenging for traditional structural determination techniques. Often, scientists have to rely on piecemeal evidence drawn from diverse experimental techniques to characterize their functional mechanisms. Multiscale simulations can help bridge critical knowledge gaps about IDP structure-function relationships-however, these techniques also face challenges in resolving emergent phenomena within IDP conformational ensembles. We posit that scalable statistical inference techniques can effectively integrate information gleaned from multiple experimental techniques as well as from simulations, thus providing access to atomistic details of these emergent phenomena.

Confronting pitfalls of AI-augmented molecular dynamics using statistical physics

Artificial intelligence (AI)-based approaches have had indubitable impact across the sciences through the ability to extract relevant information from raw data. Recently, AI has also found use in enhancing the efficiency of molecular simulations, wherein AI derived slow modes are used to accelerate the simulation in targeted ways. However, while typical fields where AI is used are characterized by a plethora of data, molecular simulations, per construction, suffer from limited sampling and thus limited data. As such, the use of AI in molecular simulations can suffer from a dangerous situation where the AI-optimization could get stuck in spurious regimes, leading to incorrect characterization of the reaction coordinate (RC) for the problem at hand. When such an incorrect RC is then used to perform additional simulations, one could start to deviate progressively from the ground truth. To deal with this problem of spurious AI-solutions, here, we report a novel and automated algorithm using ideas from statistical mechanics. It is based on the notion that a more reliable AI-solution will be one that maximizes the timescale separation between slow and fast processes. To learn this timescale separation even from limited data, we use a maximum caliber-based framework. We show the applicability of this automatic protocol for three classic benchmark problems, namely, the conformational dynamics of a model peptide, ligand-unbinding from a protein, and folding/unfolding energy landscape of the C-terminal domain of protein G. We believe that our work will lead to increased and robust use of trustworthy AI in molecular simulations of complex systems.

Thermodynamics and Kinetics of Drug-Target Binding by Molecular Simulation

Computational studies play an increasingly important role in chemistry and biophysics, mainly thanks to improvements in hardware and algorithms. In drug discovery and development, computational studies can reduce the costs and risks of bringing a new medicine to market. Computational simulations are mainly used to optimize promising new compounds by estimating their binding affinity to proteins. This is challenging due to the complexity of the simulated system. To assess the present and future value of simulation for drug discovery, we review key applications of advanced methods for sampling complex free-energy landscapes at near nonergodicity conditions and for estimating the rate coefficients of very slow processes of pharmacological interest. We outline the statistical mechanics and computational background behind this research, including methods such as steered molecular dynamics and metadynamics. We review recent applications to pharmacology and drug discovery and discuss possible guidelines for the practitioner. Recent trends in machine learning are also briefly discussed. Thanks to the rapid development of methods for characterizing and quantifying rare events, simulation's role in drug discovery is likely to expand, making it a valuable complement to experimental and clinical approaches.

AI-Driven Multiscale Simulations Illuminate Mechanisms of SARS-CoV-2 Spike Dynamics

We develop a generalizable AI-driven workflow that leverages heterogeneous HPC resources to explore the time-dependent dynamics of molecular systems. We use this workflow to investigate the mechanisms of infectivity of the SARS-CoV-2 spike protein, the main viral infection machinery. Our workflow enables more efficient investigation of spike dynamics in a variety of complex environments, including within a complete SARS-CoV-2 viral envelope simulation, which contains 305 million atoms and shows strong scaling on ORNL Summit using NAMD. We present several novel scientific discoveries, including the elucidation of the spike's full glycan shield, the role of spike glycans in modulating the infectivity of the virus, and the characterization of the flexible interactions between the spike and the human ACE2 receptor. We also demonstrate how AI can accelerate conformational sampling across different systems and pave the way for the future application of such methods to additional studies in SARS-CoV-2 and other molecular systems.

Recent progress in molecular simulation methods for drug binding kinetics

Due to the contribution of drug-target binding kinetics to drug efficacy, there is a high level of interest in developing methods to predict drug-target binding kinetic parameters. During the review period, a wide range of enhanced sampling molecular dynamics simulation-based methods has been developed for computing drug-target binding kinetics and studying binding and unbinding mechanisms. Here, we assess the performance of these methods considering two benchmark systems in detail: mutant T4 lysozyme-ligand complexes and a large set of N-HSP90-inhibitor complexes. The results indicate that some of the simulation methods can already be usefully applied in drug discovery or lead optimization programs but that further studies on more high-quality experimental benchmark datasets are necessary to improve and validate computational methods.

Capturing Protein-Ligand Recognition Pathways in Coarse-Grained Simulation

Protein-ligand recognition is dynamic and complex. A key approach in deciphering the mechanism underlying the recognition process is to capture the kinetic process of the ligand in its act of binding to its designated protein cavity. Toward this end, ultralong all-atom molecular dynamics simulation has recently emerged as a popular method of choice because of its ability to record these events at high spatial and temporal resolution. However, success via this route comes at an exorbitant computational cost. Herein, we demonstrate that coarse-grained models of the protein, when systematically optimized to maintain its tertiary fold, can capture the complete process of spontaneous protein-ligand binding from bulk media to the cavity at crystallographic precision and within wall clock time that is orders of magnitude shorter than that of all-atom simulations. The exhaustive sampling of ligand exploration in protein and solvent, harnessed by coarse-grained simulation, leads to elucidation of new ligand recognition pathways and discovery of non-native binding poses.

Machine Learning for Molecular Simulation

Machine learning (ML) is transforming all areas of science. The complex and time-consuming calculations in molecular simulations are particularly suitable for an ML revolution and have already been profoundly affected by the application of existing ML methods. Here we review recent ML methods for molecular simulation, with particular focus on (deep) neural networks for the prediction of quantum-mechanical energies and forces, on coarse-grained molecular dynamics, on the extraction of free energy surfaces and kinetics, and on generative network approaches to sample molecular equilibrium structures and compute thermodynamics. To explain these methods and illustrate open methodological problems, we review some important principles of molecular physics and describe how they can be incorporated into ML structures. Finally, we identify and describe a list of open challenges for the interface between ML and molecular simulation.

Advances in the calculation of binding free energies

In recent years, calculations of binding affinities from molecular simulations seem to have matured significantly. While the number of applications of such methods in drug design and biotechnology increases, the number of truly new methodological developments decreases. This review provides an overview of the current status of the field as reflected in recent publications. The focus is on the challenges that remain when using endstate, alchemical and pathway methods. For endstate methods this is the calculation of entropic contributions. For alchemical methods there are unsolved problems associated with the solvation of the active site, sampling slow degrees of freedom and when modifying the net charge. For pathway methods achieving sufficient sampling remains challenging. New trends are also highlighted, including the use of pathway methods for the quantification of protein-protein interactions.

Molecular Insights from Conformational Ensembles via Machine Learning

Biomolecular simulations are intrinsically high dimensional and generate noisy data sets of ever-increasing size. Extracting important features from the data is crucial for understanding the biophysical properties of molecular processes, but remains a big challenge. Machine learning (ML) provides powerful dimensionality reduction tools. However, such methods are often criticized as resembling black boxes with limited human-interpretable insight. We use methods from supervised and unsupervised ML to efficiently create interpretable maps of important features from molecular simulations. We benchmark the performance of several methods, including neural networks, random forests, and principal component analysis, using a toy model with properties reminiscent of macromolecular behavior. We then analyze three diverse biological processes: conformational changes within the soluble protein calmodulin, ligand binding to a G protein-coupled receptor, and activation of an ion channel voltage-sensor domain, unraveling features critical for signal transduction, ligand binding, and voltage sensing. This work demonstrates the usefulness of ML in understanding biomolecular states and demystifying complex simulations.

Molecular Dynamics Properties without the Full Trajectory: A Denoising Autoencoder Network for Properties of Simple Liquids

Molecular dynamics (MD) simulation is a popularly used computational tool to compute microscopic and macroscopic properties of a variety of systems including liquids, solids, biological systems, etc. To determine properties of atomic systems to a good level of accuracy with minimal noise or fluctuation, MD simulations are performed over a long time ranging from a few nanoseconds to several tens to hundreds of nanoseconds depending on the system and the properties of interest. In this study, by considering simple liquids, we explore the feasibility of significantly reducing the MD simulation time to compute various properties of monatomic systems such as the structure, pressure, and isothermal compressibility. To do so, extensive MD simulations are performed on 12 000 distinct Lennard-Jones systems at various thermodynamic states. Then, a deep denoising autoencoder network is trained to take the radial distribution function (RDF) from a single snapshot of a Lennard-Jones liquid to compute the mean, temporally averaged RDF. We show that the method is successful in the prediction of RDF and other properties such as the pressure and isothermal compressibility that can be computed based on the RDF not only for Lennard-Jones liquids at various thermodynamic states but also for various simple liquids described by exponential, Yukawa, and inverse-power-law pair potentials.

Past-future information bottleneck for sampling molecular reaction coordinate simultaneously with thermodynamics and kinetics

The ability to rapidly learn from high-dimensional data to make reliable bets about the future is crucial in many contexts. This could be a fly avoiding predators, or the retina processing gigabytes of data to guide human actions. In this work we draw parallels between these and the efficient sampling of biomolecules with hundreds of thousands of atoms. For this we use the Predictive Information Bottleneck framework used for the first two problems, and re-formulate it for the sampling of biomolecules, especially when plagued with rare events. Our method uses a deep neural network to learn the minimally complex yet most predictive aspects of a given biomolecular trajectory. This information is used to perform iteratively biased simulations that enhance the sampling and directly obtain associated thermodynamic and kinetic information. We demonstrate the method on two test-pieces, studying processes slower than milliseconds, calculating free energies, kinetics and critical mutations.

Efficient sampling of rare events in all-atom molecular dynamics simulations remains a challenge. Here, the authors adapt the Predictive Information Bottleneck framework to sample biomolecular structure and dynamics through iterative rounds of biased simulations and deep learning.

Nonlinear discovery of slow molecular modes using state-free reversible VAMPnets

The success of enhanced sampling molecular simulations that accelerate along collective variables (CVs) is predicated on the availability of variables coincident with the slow collective motions governing the long-time conformational dynamics of a system. It is challenging to intuit these slow CVs for all but the simplest molecular systems, and their data-driven discovery directly from molecular simulation trajectories has been a central focus of the molecular simulation community to both unveil the important physical mechanisms and drive enhanced sampling. In this work, we introduce state-free reversible VAMPnets (SRV) as a deep learning architecture that learns nonlinear CV approximants to the leading slow eigenfunctions of the spectral decomposition of the transfer operator that evolves equilibrium-scaled probability distributions through time. Orthogonality of the learned CVs is naturally imposed within network training without added regularization. The CVs are inherently explicit and differentiable functions of the input coordinates making them well-suited to use in enhanced sampling calculations. We demonstrate the utility of SRVs in capturing parsimonious nonlinear representations of complex system dynamics in applications to 1D and 2D toy systems where the true eigenfunctions are exactly calculable and to molecular dynamics simulations of alanine dipeptide and the WW domain protein.

Kinetics of Ligand-Protein Dissociation from All-Atom Simulations: Are We There Yet?

Large parallel gains in the development of both computational resources and sampling methods have now made it possible to simulate dissociation events in ligand-protein complexes with all-atom resolution. Such encouraging progress, together with the inherent spatiotemporal resolution associated with molecular simulations, has left their use for investigating dissociation processes brimming with potential, both in rational drug design, where it can be an invaluable tool for determining the mechanistic driving forces behind dissociation rate constants, and in force-field development, where it can provide a catalog of transient molecular structures with which to refine force fields. Although much progress has been made in making force fields more accurate, reducing their error for transient structures along a transition path could yet prove to be a critical development helping to make kinetic predictions much more accurate. In what follows, we will provide a state-of-the-art compilation of the enhanced sampling methods based on molecular dynamics (MD) simulations used to investigate the kinetics and mechanisms of ligand-protein dissociation processes. Due to the time scales of such processes being slower than what is accessible using straightforward MD simulations, several ingenious schemes are being devised at a rapid rate to overcome this obstacle. Here we provide an up-to-date compendium of such methods and their achievements and shortcomings in extracting mechanistic insight into ligand-protein dissociation. We conclude with a critical and provocative appraisal attempting to answer the title of this Perspective.