Accurate Prediction Of Binding Affinity Research Articles

FEP Augmentation as a Means to Solve Data Paucity Problems for Machine Learning in Chemical Biology.

In the realm of medicinal chemistry, the primary objective is to swiftly optimize a multitude of chemical properties of a set of compounds to yield a clinical candidate poised for clinical trials. In recent years, two computational techniques, machine learning (ML) and physics-based methods, have evolved substantially and are now frequently incorporated into the medicinal chemist's toolbox to enhance the efficiency of both hit optimization and candidate design. Both computational methods come with their own set of limitations, and they are often used independently of each other. ML's capability to screen extensive compound libraries expediently is tempered by its reliance on quality data, which can be scarce especially during early-stage optimization. Contrarily, physics-based approaches like free energy perturbation (FEP) are frequently constrained by low throughput and high cost by comparison; however, physics-based methods are capable of making highly accurate binding affinity predictions. In this study, we harnessed the strength of FEP to overcome data paucity in ML by generating virtual activity data sets which then inform the training of algorithms. Here, we show that ML algorithms trained with an FEP-augmented data set could achieve comparable predictive accuracy to data sets trained on experimental data from biological assays. Throughout the paper, we emphasize key mechanistic considerations that must be taken into account when aiming to augment data sets and lay the groundwork for successful implementation. Ultimately, the study advocates for the synergy of physics-based methods and ML to expedite the lead optimization process. We believe that the physics-based augmentation of ML will significantly benefit drug discovery, as these techniques continue to evolve.

Evaluation of Cx43 Gap Junction Inhibitors Using a Quantitative Structure-Activity Relationship Model.

Gap junctions (GJs) made of connexin-43 (Cx43) are necessary for the conduction of electrical impulses in the heart. Modulation of Cx43 GJ activity may be beneficial in the treatment of cardiac arrhythmias and other dysfunctions. The search for novel GJ-modulating agents using molecular docking allows for the accurate prediction of binding affinities of ligands, which, unfortunately, often poorly correlate with their potencies. The objective of this study was to demonstrate that a Quantitative Structure-Activity Relationship (QSAR) model could be used for more precise identification of potent Cx43 GJ inhibitors. Using molecular docking, QSAR, and 3D-QSAR, we evaluated 16 known Cx43 GJ inhibitors, suggested the monocyclic monoterpene d-limonene as a putative Cx43 inhibitor, and tested it experimentally in HeLa cells expressing exogenous Cx43. The predicted concentrations required to produce 50% of the maximal effect (IC50) for each of these compounds were compared with those determined experimentally (pIC50 and eIC50, respectively). The pIC50ies of d-limonene and other Cx43 GJ inhibitors examined by our QSAR and 3D-QSAR models showed a good correlation with their eIC50ies (R = 0.88 and 0.90, respectively) in contrast to pIC50ies obtained from molecular docking (R = 0.78). However, molecular docking suggests that inhibitor potency may depend on their docking conformation on Cx43. Searching for new potent, selective, and specific inhibitors of GJ channels, we propose to perform the primary screening of new putative compounds using the QSAR model, followed by the validation of the most suitable candidates by patch-clamp techniques.

SophosQM: Accurate Binding Affinity Prediction in Compound Optimization.

The optimization of compounds' binding affinity for a biological target is a crucial aspect of the drug development process. Being able to accurately predict binding energies in advance of synthesizing compounds would have a massive impact on the speed of the drug discovery process. The ideal binding affinity prediction method should combine accuracy, reliability, and speed. In this paper, we present SophosQM, a quantum mechanics (QM)-based approach, which can accurately predict the binding affinities of compounds to proteins. The binding affinity predictive models generated by SophosQM are based on the fragment molecular orbital (FMO) method to estimate the enthalpic component of the binding free energy, and a macroscopic descriptor, clog P, is used as an approximation of the entropic component. The affinity prediction is performed using multilinear regression, fitting the experimental values against the FMO-computed enthalpic term and clog P. The quality of the prediction can be assessed in terms of the correlation coefficient between experimental and predicted values. In this work, the method's reliability and accuracy are exemplified by applying SophosQM to 70 compounds binding to six different targets of pharmaceutical relevance. Overall, the results show a very satisfactory performance with a global correlation coefficient in the order of 0.9. Our predictions also show a satisfactory performance compared to data based on free energy perturbation. Finally, SophosQM can also be applied in high-throughput mode by using semiempirical QM methods to evaluate large portions of chemical space, while retaining a good level of accuracy, but decreasing the computing time to just a few seconds per compound.

Multi-shelled ECIF: improved extended connectivity interaction features for accurate binding affinity prediction

Abstract Motivation Extended connectivity interaction features (ECIF) is a method developed to predict protein–ligand binding affinity, allowing for detailed atomic representation. It performed very well in terms of Comparative Assessment of Scoring Functions 2016 (CASF-2016) scoring power. However, ECIF has the limitation of not being able to adequately account for interatomic distances. Results To investigate what kind of distance representation is effective for P-L binding affinity prediction, we have developed two algorithms that improved ECIF’s feature extraction method to take distance into account. One is multi-shelled ECIF, which takes into account the distance between atoms by dividing the distance between atoms into multiple layers. The other is weighted ECIF, which weights the importance of interactions according to the distance between atoms. A comparison of these two methods shows that multi-shelled ECIF outperforms weighted ECIF and the original ECIF, achieving a CASF-2016 scoring power Pearson correlation coefficient of 0.877. Availability and implementation All the codes and data are available on GitHub (https://github.com/koji11235/MSECIFv2).

Ligand binding affinity prediction with fusion of graph neural networks and 3D structure-based complex graph.

Accurate prediction of protein-ligand binding affinity is pivotal for drug design and discovery. Here, we proposed a novel deep fusion graph neural networks framework named FGNN to learn the protein-ligand interactions from the 3D structures of protein-ligand complexes. Unlike 1D sequences for proteins or 2D graphs for ligands, the 3D graph of protein-ligand complex enables the more accurate representations of the protein-ligand interactions. Benchmark studies have shown that our fusion models FGNN can achieve more accurate prediction of binding affinity than any individual algorithm. The advantages of fusion strategies have been demonstrated in terms of expressive power of data, learning efficiency and model interpretability. Our fusion models show satisfactory performances on diverse data sets, demonstrating their generalization ability. Given the good performances in both binding affinity prediction and virtual screening, our fusion models are expected to be practically applied for drug screening and design. Our work highlights the potential of the fusion graph neural network algorithm in solving complex prediction problems in computational biology and chemistry. The fusion graph neural networks (FGNN) model is freely available in https://github.com/LinaDongXMU/FGNN.

PLAS-5k: Dataset of Protein-Ligand Affinities from Molecular Dynamics for Machine Learning Applications

Computational methods and recently modern machine learning methods have played a key role in structure-based drug design. Though several benchmarking datasets are available for machine learning applications in virtual screening, accurate prediction of binding affinity for a protein-ligand complex remains a major challenge. New datasets that allow for the development of models for predicting binding affinities better than the state-of-the-art scoring functions are important. For the first time, we have developed a dataset, PLAS-5k comprised of 5000 protein-ligand complexes chosen from PDB database. The dataset consists of binding affinities along with energy components like electrostatic, van der Waals, polar and non-polar solvation energy calculated from molecular dynamics simulations using MMPBSA (Molecular Mechanics Poisson-Boltzmann Surface Area) method. The calculated binding affinities outperformed docking scores and showed a good correlation with the available experimental values. The availability of energy components may enable optimization of desired components during machine learning-based drug design. Further, OnionNet model has been retrained on PLAS-5k dataset and is provided as a baseline for the prediction of binding affinities.

Machine-Learning- and Knowledge-Based Scoring Functions Incorporating Ligand and Protein Fingerprints.

We propose a novel machine-learning-based scoring function for drug discovery that incorporates ligand and protein structural information into a knowledge-based PMF score. Molecular docking, a simulation method for structure-based drug design (SBDD), is expected to reduce the enormous costs associated with conventional experimental methods in terms of rational drug discovery. Molecular docking has two main purposes: to predict ligand-binding structures for target proteins and to predict protein–ligand binding affinity. Currently available programs of molecular docking offer an accurate prediction of ligand binding structures for many systems. However, the accurate prediction of binding affinity remains challenging. In this study, we developed a new scoring function that incorporates fingerprints representing ligand and protein structures as descriptors in the PMF score. Here, regression analysis of the scoring function was performed using the following machine learning techniques: least absolute shrinkage and selection operator (LASSO) and light gradient boosting machine (LightGBM). The results on a test data set showed that the binding affinity delivered by the newly developed scoring function has a Pearson correlation coefficient of 0.79 with the experimental value, which surpasses that of the conventional scoring functions. Further analysis provided a chemical understanding of the descriptors that contributed significantly to the improvement in prediction accuracy. Our approach and findings are useful for rational drug discovery.

On the Frustration to Predict Binding Affinities from Protein-Ligand Structures with Deep Neural Networks.

Accurate prediction of binding affinities from protein-ligand atomic coordinates remains a major challenge in early stages of drug discovery. Using modular message passing graph neural networks describing both the ligand and the protein in their free and bound states, we unambiguously evidence that an explicit description of protein-ligand noncovalent interactions does not provide any advantage with respect to ligand or protein descriptors. Simple models, inferring binding affinities of test samples from that of the closest ligands or proteins in the training set, already exhibit good performances, suggesting that memorization largely dominates true learning in the deep neural networks. The current study suggests considering only noncovalent interactions while omitting their protein and ligand atomic environments. Removing all hidden biases probably requires much denser protein-ligand training matrices and a coordinated effort of the drug design community to solve the necessary protein-ligand structures.

The Impact of Experimental and Calculated Error on the Performance of Affinity Predictions.

The accurate prediction of binding affinity between protein and small molecules with free energy methods, particularly the difference in binding affinities via relative binding free energy calculations, has undergone a dramatic increase in use and impact over recent years. The improvements in methodology, hardware, and implementation can deliver results with less than 1 kcal/mol mean unsigned error between calculation and experiment. This is a remarkable achievement and beckons some reflection on the significance of calculation approaching the accuracy of experiment. In this article, we describe a statistical analysis of the implications of variance (standard deviation) of both experimental and calculated binding affinities with respect to the unknown true binding affinity. We reveal that plausible ratios of standard deviation in experiment and calculation can lead to unexpected outcomes for assessing the performance of predictions. The work extends beyond the case of binding free energies to other affinity or property prediction methods.

A Cascade Graph Convolutional Network for Predicting Protein-Ligand Binding Affinity.

Accurate prediction of binding affinity between protein and ligand is a very important step in the field of drug discovery. Although there are many methods based on different assumptions and rules do exist, prediction performance of protein–ligand binding affinity is not satisfactory so far. This paper proposes a new cascade graph-based convolutional neural network architecture by dealing with non-Euclidean irregular data. We represent the molecule as a graph, and use a simple linear transformation to deal with the sparsity problem of the one-hot encoding of original data. The first stage adopts ARMA graph convolutional neural network to learn the characteristics of atomic space in the protein–ligand complex. In the second stage, one variant of the MPNN graph convolutional neural network is introduced with chemical bond information and interactive atomic features. Finally, the architecture passes through the global add pool and the fully connected layer, and outputs a constant value as the predicted binding affinity. Experiments on the PDBbind v2016 data set showed that our method is better than most of the current methods. Our method is also comparable to the state-of-the-art method on the data set, and is more intuitive and simple.

Exploring the bioactivity of pentacyclic triterpenoids as potential antimycobacterial nutraceutics: Insights through comparative biomolecular modelling

A group of bioactive compounds known as triterpenoids, which are often found in plant materials, have been tested to possess nutritional and pharmaceutical activity. These plant components are referred to as nutraceuticals, and are used as therapeutic agents. In this study, we explore the interactions of betulinic acid (BA), oleanolic acid (OA), ursolic acid (UA), and maslinic acid (MA) against FadA5. Studies have identified FadA5, a trifunctional enzyme-like thiolase, as a target towards Mycobacterium tuberculosis inhibition. The investigation involves molecular dynamics (MD) and hybrid quantum mechanics/molecular mechanics (QM/MM) applications. Analyses of the four pentacyclic triterpenoids binding to FadA5 showed appreciable bioactivity against FadA5. The application of two or more theoretical models to unravel ligand—enzyme binding energies can pave the way for accurate binding affinity prediction and validation.

New machine learning and physics-based scoring functions for drug discovery

Scoring functions are essential for modern in silico drug discovery. However, the accurate prediction of binding affinity by scoring functions remains a challenging task. The performance of scoring functions is very heterogeneous across different target classes. Scoring functions based on precise physics-based descriptors better representing protein–ligand recognition process are strongly needed. We developed a set of new empirical scoring functions, named DockTScore, by explicitly accounting for physics-based terms combined with machine learning. Target-specific scoring functions were developed for two important drug targets, proteases and protein–protein interactions, representing an original class of molecules for drug discovery. Multiple linear regression (MLR), support vector machine and random forest algorithms were employed to derive general and target-specific scoring functions involving optimized MMFF94S force-field terms, solvation and lipophilic interactions terms, and an improved term accounting for ligand torsional entropy contribution to ligand binding. DockTScore scoring functions demonstrated to be competitive with the current best-evaluated scoring functions in terms of binding energy prediction and ranking on four DUD-E datasets and will be useful for in silico drug design for diverse proteins as well as for specific targets such as proteases and protein–protein interactions. Currently, the MLR DockTScore is available at www.dockthor.lncc.br.

Structure-based protein–ligand interaction fingerprints for binding affinity prediction

Binding affinity prediction (BAP) using protein-ligand complex structures is crucial to computer-aided drug design, but remains a challenging problem. To achieve efficient and accurate BAP, machine-learning scoring functions (SFs) based on a wide range of descriptors have been developed. Among those descriptors, protein-ligand interaction fingerprints (IFPs) are competitive due to their simple representations, elaborate profiles of key interactions and easy collaborations with machine-learning algorithms. In this paper, we have adopted a building-block-based taxonomy to review a broad range of IFP models, and compared representative IFP-based SFs in target-specific and generic scoring tasks. Atom-pair-counts-based and substructure-based IFPs show great potential in these tasks.

DockingApp RF: A State-of-the-Art Novel Scoring Function for Molecular Docking in a User-Friendly Interface to AutoDock Vina.

Motivation: Bringing a new drug to the market is expensive and time-consuming. To cut the costs and time, computer-aided drug design (CADD) approaches have been increasingly included in the drug discovery pipeline. However, despite traditional docking tools show a good conformational space sampling ability, they are still unable to produce accurate binding affinity predictions. This work presents a novel scoring function for molecular docking seamlessly integrated into DockingApp, a user-friendly graphical interface for AutoDock Vina. The proposed function is based on a random forest model and a selection of specific features to overcome the existing limits of Vina’s original scoring mechanism. A novel version of DockingApp, named DockingApp RF, has been developed to host the proposed scoring function and to automatize the rescoring procedure of the output of AutoDock Vina, even to nonexpert users. Results: By coupling intermolecular interaction, solvent accessible surface area features and Vina’s energy terms, DockingApp RF’s new scoring function is able to improve the binding affinity prediction of AutoDock Vina. Furthermore, comparison tests carried out on the CASF-2013 and CASF-2016 datasets demonstrate that DockingApp RF’s performance is comparable to other state-of-the-art machine-learning- and deep-learning-based scoring functions. The new scoring function thus represents a significant advancement in terms of the reliability and effectiveness of docking compared to AutoDock Vina’s scoring function. At the same time, the characteristics that made DockingApp appealing to a wide range of users are retained in this new version and have been complemented with additional features.

Rapid, accurate, precise and reproducible ligand-protein binding free energy prediction.

A central quantity of interest in molecular biology and medicine is the free energy of binding of a molecule to a target biomacromolecule. Until recently, the accurate prediction of binding affinity had been widely regarded as out of reach of theoretical methods owing to the lack of reproducibility of the available methods, not to mention their complexity, computational cost and time-consuming procedures. The lack of reproducibility stems primarily from the chaotic nature of classical molecular dynamics (MD) and the associated extreme sensitivity of trajectories to their initial conditions. Here, we review computational approaches for both relative and absolute binding free energy calculations, and illustrate their application to a diverse set of ligands bound to a range of proteins with immediate relevance in a number of medical domains. We focus on ensemble-based methods which are essential in order to compute statistically robust results, including two we have recently developed, namely thermodynamic integration with enhanced sampling and enhanced sampling of MD with an approximation of continuum solvent. Together, these form a set of rapid, accurate, precise and reproducible free energy methods. They can be used in real-world problems such as hit-to-lead and lead optimization stages in drug discovery, and in personalized medicine. These applications show that individual binding affinities equipped with uncertainty quantification may be computed in a few hours on a massive scale given access to suitable high-end computing resources and workflow automation. A high level of accuracy can be achieved using these approaches.

Analysis of Software Methods for Estimation of Protein-Protein Relative Binding Affinity

Analysis of Software Methods for Estimation of Protein-Protein Relative Binding Affinity: Biophysical modeling of protein-protein interactions provides insight into how and why proteins behave in specific ways and is useful for estimating how amino acid mutations modify protein-protein binding affinity, a topic with significant clinical applications. Binding affinity prediction software vary in the complexity of information used to create predictions. Our hypothesis is that software methods using a wider variety of information will provide more accurate binding affinity predictions than those relying on a single descriptive energy function. We compare six methods that range from empirical to semi-empirical. We generated estimates for sixteen protein complex test systems with experimental data. A performance score for each program was determined based on correlation to experimental data and the ability to correctly estimate the sign of the relative binding affinity.

Comparison of affinity ranking using AutoDock-GPU and MM-GBSA scores for BACE-1 inhibitors in the D3R Grand Challenge 4.

Molecular docking has been successfully used in computer-aided molecular design projects for the identification of ligand poses within protein binding sites. However, relying on docking scores to rank different ligands with respect to their experimental affinities might not be sufficient. It is believed that the binding scores calculated using molecular mechanics combined with the Poisson-Boltzman surface area (MM-PBSA) or generalized Born surface area (MM-GBSA) can predict binding affinitiesmore accurately. In this perspective, we decided to take part in Stage 2 of the Drug Design Data Resource (D3R) Grand Challenge 4 (GC4) to compare the performance of a quick scoring function, AutoDock4, to that of MM-GBSA in predicting the binding affinities of a set of [Formula: see text]-Amyloid Cleaving Enzyme 1 (BACE-1) ligands. Our results show that re-scoring docking poses using MM-GBSA did not improve the correlation with experimental affinities. We further did a retrospective analysis of the results and found that our MM-GBSA protocol is sensitive to details in the protein-ligand system: (i) neutral ligands are more adapted to MM-GBSA calculations than charged ligands, (ii) predicted binding affinities depend on the initial conformation of the BACE-1 receptor, (iii) protonating the aspartyl dyad of BACE-1 correctly results in more accurate binding affinity predictions.

Assessing the Performance of MM/PBSA, MM/GBSA, and QM-MM/GBSA Approaches on Protein/Carbohydrate Complexes: Effect of Implicit Solvent Models, QM Methods, and Entropic Contributions.

Rapid and accurate binding affinity prediction of protein-carbohydrate complexes is a major challenge in glycomimetics design. Among the existing computational techniques, end-point methods have received considerable interest because of their low computational cost. However, significant obstacles remain when such methods are applied to protein-glycan complexes. This article reports the performance of end-point free-energy calculation methods: molecular mechanics/Poisson-Boltzmann surface area (MM/PBSA), MM/generalized Born surface area (MM/GBSA), and quantum mechanics-MM/GBSA (QM-MM/GBSA) on monosaccharides bound to RSL lectin from Ralstonia solanacearum. A careful investigation of the molecular dynamics simulation length, van der Waals radii sets, GB models, QM Hamiltonians, and entropic compensation has been made, and the results are compared with the experimental binding free energies from isothermal titration calorimetry/surface plasmon resonance measurements. The binding free energies using implicit solvent methods are found to be sensitive to the simulation length, radii set, GB model, and QM Hamiltonian. A simulation length of 10 ns using the radii set mbondi provides the best agreement with the experimental values ( r2 = 0.96) by MM/PBSA. The GBHCT model is in accord with the experimental values in MM/GBSA ( r2 = 0.91) or in combination with parameterized model number 6 (PM6) ( r2 = 0.98) in QM-MM/GBSA. Out of 12 QM Hamiltonians tested, PM6, density functional theory-based tight binding (DFTB), and their variants proved to be more efficient than other semiempirical methods. These methods perform equally well in predicting both absolute and relative binding free energies.

Binding affinity prediction of nanobody-protein complexes by scoring of molecular dynamics trajectories.

Nanobodies offer a viable alternative to antibodies for engineering high affinity binders. Their small size has an additional advantage: it allows exploiting computational protocols for optimizing their biophysical features, such as the binding affinity. The efficient prediction of this quantity is still considered a daunting task especially for modelled complexes. We show how molecular dynamics can successfully assist in the binding affinity prediction of modelled nanobody-protein complexes. The approximate initial configurations obtained by in silico design must undergo large rearrangements before achieving a stable conformation, in which the binding affinity can be meaningfully estimated. The scoring functions developed for the affinity evaluation of crystal structures will provide accurate estimates for modelled binding complexes if the scores are averaged over long finite temperature molecular dynamics simulations.

Toward Fast and Accurate Binding Affinity Prediction with pmemdGTI: An Efficient Implementation of GPU-Accelerated Thermodynamic Integration.

We report the implementation of the thermodynamic integration method on the pmemd module of the AMBER 16 package on GPUs (pmemdGTI). The pmemdGTI code typically delivers over 2 orders of magnitude of speed-up relative to a single CPU core for the calculation of ligand-protein binding affinities with no statistically significant numerical differences and thus provides a powerful new tool for drug discovery applications.