Leveraging nonstructural data to predict structures and affinities of protein-ligand complexes

Fully Flexible Molecular Alignment Enables Accurate Ligand Structure Modeling

Accurate protein-ligand binding poses are the prerequisites of structure-based binding affinity prediction and provide the structural basis for in-depth lead optimization in small molecule drug design. However, it is challenging to provide reasonable predictions of binding poses for different molecules due to the complexity and diversity of the chemical space of small molecules. Similarity-based molecular alignment techniques can effectively narrow the search range, as structurally similar molecules are likely to have similar binding modes, with higher similarity usually correlated to higher success rates. However, molecular similarity is not consistently high because molecules often require changes to achieve specific purposes, leading to reduced alignment precision. To address this issue, we propose a new alignment method─Z-align. This method uses topological structural information as a criterion for evaluating similarity, reducing the reliance on molecular fingerprint similarity. Our method has achieved success rates significantly higher than those of other methods at moderate levels of similarity. Additionally, our approach can comprehensively and flexibly optimize bond lengths and angles of molecules, maintaining a high accuracy even when dealing with larger molecules. Consequently, our proposed solution helps in achieving more accurate binding poses in protein-ligand docking problems, facilitating the development of small molecule drugs. Z-align is freely available as a web server at https://cloud.zelixir.com/zalign/home.

Reducing overconfident errors in molecular property classification using Posterior Network

Deep-learning-based classification models are increasingly used for predicting molecular properties in drug development. However, traditional classification models using the Softmax function often give overconfident mispredictions for out-of-distribution samples, highlighting a critical lack of accurate uncertainty estimation. Such limitations can result in substantial costs and should be avoided during drug development. Inspired by advances in evidential deep learning and Posterior Network, we replaced the Softmax function with a normalizing flow to enhance the uncertainty estimation ability of the model in molecular property classification. The proposed strategy was evaluated across diverse scenarios, including simulated experiments based on a synthetic dataset, ADMET predictions, and ligand-based virtual screening. The results demonstrate that compared with the vanilla model, the proposed strategy effectively alleviates the problem of giving overconfident but incorrect predictions. Our findings support the promising application of evidential deep learning in drug development and offer a valuable framework for further research.

A High-Quality Data Set of Protein-Ligand Binding Interactions Via Comparative Complex Structure Modeling

High-quality protein-ligand complex structures provide the basis for understanding the nature of noncovalent binding interactions at the atomic level and enable structure-based drug design. However, experimentally determined complex structures are scarce compared with the vast chemical space. In this study, we addressed this issue by constructing the BindingNet data set via comparative complex structure modeling, which contains 69,816 modeled high-quality protein-ligand complex structures with experimental binding affinity data. BindingNet provides valuable insights into investigating protein-ligand interactions, allowing visual inspection and interpretation of structural analogues' structure-activity relationships. It can also be used for evaluating machine-learning-based scoring functions. Our results indicate that machine learning models trained on BindingNet could reduce the bias caused by buried solvent-accessible surface area, as we previously found for models trained on the PDBbind data set. We also discussed strategies to improve BindingNet and its potential utilization for benchmarking the molecular docking methods and ligand binding free energy calculation approaches. The BindingNet complements PDBbind in constructing a sufficient and unbiased protein-ligand binding data set and is freely available at http://bindingnet.huanglab.org.cn.

CarsiDock: a deep learning paradigm for accurate protein-ligand docking and screening based on large-scale pre-training

The expertise accumulated in deep neural network-based structure prediction has been widely transferred to the field of protein-ligand binding pose prediction, thus leading to the emergence of a variety of deep learning-guided docking models for predicting protein-ligand binding poses without relying on heavy sampling. However, their prediction accuracy and applicability are still far from satisfactory, partially due to the lack of protein-ligand binding complex data. To this end, we create a large-scale complex dataset containing ∼9 M protein-ligand docking complexes for pre-training, and propose CarsiDock, the first deep learning-guided docking approach that leverages pre-training of millions of predicted protein-ligand complexes. CarsiDock contains two main stages, i.e., a deep learning model for the prediction of protein-ligand atomic distance matrices, and a translation, rotation and torsion-guided geometry optimization procedure to reconstruct the matrices into a credible binding pose. The pre-training and multiple innovative architectural designs facilitate the dramatically improved docking accuracy of our approach over the baselines in terms of multiple docking scenarios, thereby contributing to its outstanding early recognition performance in several retrospective virtual screening campaigns. Further explorations demonstrate that CarsiDock can not only guarantee the topological reliability of the binding poses but also successfully reproduce the crucial interactions in crystalized structures, highlighting its superior applicability.

Predictive Models and Impact of Interfacial Contacts and Amino Acids on Protein-Protein Binding Affinity

Protein-protein interactions (PPIs) play a central role in nearly all cellular processes. The strength of the binding in a PPI is characterized by the binding affinity (BA) and is a key factor in controlling protein-protein complex formation and defining the structure-function relationship. Despite advancements in understanding protein-protein binding, much remains unknown about the interfacial region and its association with BA. New models are needed to predict BA with improved accuracy for therapeutic design. Here, we use machine learning approaches to examine how well different types of interfacial contacts can be used to predict experimentally determined BA and to reveal the impact of the specific amino acids at the binding interface on BA. We create a series of multivariate linear regression models incorporating different contact features at both residue and atomic levels and examine how different methods of identifying and characterizing these properties impact the performance of these models. Particularly, we introduce a new and simple approach to predict BA based on the quantities of specific amino acids at the protein-protein interface. We found that the numbers of specific amino acids at the protein-protein interface were correlated with BA. We show that the interfacial numbers of amino acids can be used to produce models with consistently good performance across different data sets, indicating the importance of the identities of interfacial amino acids in underlying BA. When trained on a diverse set of complexes from two benchmark data sets, the best performing BA model was generated with an explicit linear equation involving six amino acids. Tyrosine, in particular, was identified as the key amino acid in controlling BA, as it had the strongest correlation with BA and was consistently identified as the most important amino acid in feature importance studies. Glycine and serine were identified as the next two most important amino acids in predicting BA. The results from this study further our understanding of PPIs and can be used to make improved predictions of BA, giving them implications for drug design and screening in the pharmaceutical industry.

How accurately can one predict drug binding modes using AlphaFold models?

Computational prediction of protein structure has been pursued intensely for decades, motivated largely by the goal of using structural models for drug discovery. Recently developed machine-learning methods such as AlphaFold 2 (AF2) have dramatically improved protein structure prediction, with reported accuracy approaching that of experimentally determined structures. To what extent do these advances translate to an ability to predict more accurately how drugs and drug candidates bind to their target proteins? Here, we carefully examine the utility of AF2 protein structure models for predicting binding poses of drug-like molecules at the largest class of drug targets, the G-protein-coupled receptors. We find that AF2 models capture binding pocket structures much more accurately than traditional homology models, with errors nearly as small as differences between structures of the same protein determined experimentally with different ligands bound. Strikingly, however, the accuracy of ligand-binding poses predicted by computational docking to AF2 models is not significantly higher than when docking to traditional homology models and is much lower than when docking to structures determined experimentally without these ligands bound. These results have important implications for all those who might use predicted protein structures for drug discovery.

Open-ComBind: harnessing unlabeled data for improved binding pose prediction

Determination of the bound pose of a ligand is a critical first step in many in silico drug discovery tasks. Molecular docking is the main tool for the prediction of non-covalent binding of a protein and ligand system. Molecular docking pipelines often only utilize the information of one ligand binding to the protein despite the commonly held hypothesis that different ligands share binding interactions when bound to the same receptor. Here we describe Open-ComBind, an easy-to-use, open-source version of the ComBind molecular docking pipeline that leverages information from multiple ligands without known bound structures to enhance pose selection. We first create distributions of feature similarities between ligand pose pairs, comparing near-native poses with all sampled docked poses. These distributions capture the likelihood of observing similar features, such as hydrogen bonds or hydrophobic contacts, in different pose configurations. These similarity distributions are then combined with a per-ligand docking score to enhance overall pose selection by 5% and 4.5% for high-affinity and congeneric series helper ligands, respectively. Open-ComBind reduces the average RMSD of ligands in our benchmark dataset by 9.0%. We provide Open-ComBind as an easy-to-use command line and Python API to increase pose prediction performance at www.github.com/drewnutt/open_combind .

Enhancing Ligand Pose Sampling for Molecular Docking

Deep learning promises to dramatically improve scoring functions for molecular docking, leading to substantial advances in binding pose prediction and virtual screening. To train scoring functions-and to perform molecular docking-one must generate a set of candidate ligand binding poses. Unfortunately, the sampling protocols currently used to generate candidate poses frequently fail to produce any poses close to the correct, experimentally determined pose, unless information about the correct pose is provided. This limits the accuracy of learned scoring functions and molecular docking. Here, we describe two improved protocols for pose sampling: GLOW (auGmented sampLing with sOftened vdW potential) and a novel technique named IVES (IteratiVe Ensemble Sampling). Our benchmarking results demonstrate the effectiveness of our methods in improving the likelihood of sampling accurate poses, especially for binding pockets whose shape changes substantially when different ligands bind. This improvement is observed across both experimentally determined and AlphaFold-generated protein structures. Additionally, we present datasets of candidate ligand poses generated using our methods for each of around 5,000 protein-ligand cross-docking pairs, for training and testing scoring functions. To benefit the research community, we provide these cross-docking datasets and an open-source Python implementation of GLOW and IVES at https://github.com/drorlab/GLOW_IVES.

Binding affinity predictions with hybrid quantum-classical convolutional neural networks

Central in drug design is the identification of biomolecules that uniquely and robustly bind to a target protein, while minimizing their interactions with others. Accordingly, precise binding affinity prediction, enabling the accurate selection of suitable candidates from an extensive pool of potential compounds, can greatly reduce the expenses associated to practical experimental protocols. In this respect, recent advances revealed that deep learning methods show superior performance compared to other traditional computational methods, especially with the advent of large datasets. These methods, however, are complex and very time-intensive, thus representing an important clear bottleneck for their development and practical application. In this context, the emerging realm of quantum machine learning holds promise for enhancing numerous classical machine learning algorithms. In this work, we take one step forward and present a hybrid quantum-classical convolutional neural network, which is able to reduce by 20% the complexity of the classical counterpart while still maintaining optimal performance in the predictions. Additionally, this results in a significant cost and time savings of up to 40% in the training stage, which means a substantial speed-up of the drug design process.

Computing the relative binding affinity of ligands based on a pairwise binding comparison network

Structure-based lead optimization is an open challenge in drug discovery, which is still largely driven by hypotheses and depends on the experience of medicinal chemists. Here we propose a pairwise binding comparison network (PBCNet) based on a physics-informed graph attention mechanism, specifically tailored for ranking the relative binding affinity among congeneric ligands. Benchmarking on two held-out sets (provided by Schrödinger and Merck) containing over 460 ligands and 16 targets, PBCNet demonstrated substantial advantages in terms of both prediction accuracy and computational efficiency. Equipped with a fine-tuning operation, the performance of PBCNet reaches that of Schrödinger's FEP+, which is much more computationally intensive and requires substantial expert intervention. A further simulation-based experiment showed that active learning-optimized PBCNet may accelerate lead optimization campaigns by 473%. Finally, for the convenience of users, a web service for PBCNet is established to facilitate complex relative binding affinity prediction through an easy-to-operate graphical interface.

PrePCI: A structure- and chemical similarity-informed database of predicted protein compound interactions

We describe the Predicting Protein-Compound Interactions (PrePCI) database which comprises over 5 billion predicted interactions between 6.8 million chemical compounds and 19,797 human proteins. PrePCI relies on a proteome-wide database of structural models based on both traditional modeling techniques and the AlphaFold Protein Structure Database. Sequence- and structural similarity-based metrics are established between template proteins, T, in the Protein Data Bank that bind compounds, C, and query proteins in the model database, Q. When the metrics exceed threshold values, it is assumed that C also binds to Q with a likelihood ratio (LR) derived from machine learning. If the relationship is based on structural similarity, the LR is based on a scoring function that measures the extent to which C is compatible with the binding site of Q as described in the LT-scanner algorithm. For every predicted complex derived in this way, chemical similarity based on the Tanimoto coefficient identifies other small molecules that may bind to Q. An overall LR for the binding of C to Q is obtained from Naive Bayesian statistics. The PrePCI database can be queried by entering a UniProt ID or gene name for a protein to obtain a list of compounds predicted to bind to it along with associated LRs. Alternatively, entering an identifier for the compound outputs a list of proteins it is predicted to bind. Specific applications of the database to lead discovery, elucidation of drug mechanism of action, and biological function annotation are described.

The Current Perspectives in Clinical Research: Computer-Assisted Drug Designing, Ethics, and Good Clinical Practice

In the era of emerging microbial and non-communicable diseases and re-emerging microbial infections, the medical fraternity and the public are plagued by under-preparedness. It is evident by the severity of the Coronavirus disease (COVID-19) pandemic that novel microbial diseases are a challenge and are challenging to control. This is mainly attributed to the lack of complete knowledge of the novel microbe’s biology and pathogenesis and the unavailability of therapeutic drugs and vaccines to treat and control the disease. Clinical research is the only answer utilizing which can handle most of these circumstances. In this review, we highlight the importance of computer-assisted drug designing (CADD) and the aspects of molecular docking, molecular superimposition, 3D-pharmacophore technology, ethics, and good clinical practice (GCP) for the development of therapeutic drugs, devices, and vaccines.

CB-Dock2: improved protein-ligand blind docking by integrating cavity detection, docking and homologous template fitting

Protein-ligand blind docking is a powerful method for exploring the binding sites of receptors and the corresponding binding poses of ligands. It has seen wide applications in pharmaceutical and biological researches. Previously, we proposed a blind docking server, CB-Dock, which has been under heavy use (over 200 submissions per day) by researchers worldwide since 2019. Here, we substantially improved the docking method by combining CB-Dock with our template-based docking engine to enhance the accuracy in binding site identification and binding pose prediction. In the benchmark tests, it yielded the success rate of ∼85% for binding pose prediction (RMSD < 2.0 Å), which outperformed original CB-Dock and most popular blind docking tools. This updated docking server, named CB-Dock2, reconfigured the input and output web interfaces, together with a highly automatic docking pipeline, making it a particularly efficient and easy-to-use tool for the bioinformatics and cheminformatics communities. The web server is freely available at https://cadd.labshare.cn/cb-dock2/.

FitDock: protein-ligand docking by template fitting

Protein-ligand docking is an essential method in computer-aided drug design and structural bioinformatics. It can be used to identify active compounds and reveal molecular mechanisms of biological processes. A successful docking usually requires thorough conformation sampling and scoring, which are computationally expensive and difficult. Recent studies demonstrated that it can be beneficial to docking with the guidance of existing similar co-crystal structures. In this work, we developed a protein-ligand docking method, named FitDock, which fits initial conformation to the given template using a hierarchical multi-feature alignment approach, subsequently explores the possible conformations and finally outputs refined docking poses. In our comprehensive benchmark tests, FitDock showed 40%-60% improvement in terms of docking success rate and an order of magnitude faster over popular docking methods, if template structures exist (> 0.5 ligand similarity). FitDock has been implemented in a user-friendly program, which could serve as a convenient tool for drug design and molecular mechanism exploration. It is now freely available for academic users at http://cao.labshare.cn/fitdock/.