Prediction Of Protein-ligand Interactions Research Articles

A new paradigm for applying deep learning to protein-ligand interaction prediction.

Protein-ligand interaction prediction presents a significant challenge in drug design. Numerous machine learning and deep learning (DL) models have been developed to accurately identify docking poses of ligands and active compounds against specific targets. However, current models often suffer from inadequate accuracy or lack practical physical significance in their scoring systems. In this research paper, we introduce IGModel, a novel approach that utilizes the geometric information of protein-ligand complexes as input for predicting the root mean square deviation of docking poses and the binding strength (pKd, the negative value of the logarithm of binding affinity) within the same prediction framework. This ensures that the output scores carry intuitive meaning. We extensively evaluate the performance of IGModel on various docking power test sets, including the CASF-2016 benchmark, PDBbind-CrossDocked-Core and DISCO set, consistently achieving state-of-the-art accuracies. Furthermore, we assess IGModel's generalizability and robustness by evaluating it on unbiased test sets and sets containing target structures generated by AlphaFold2. The exceptional performance of IGModel on these sets demonstrates its efficacy. Additionally, we visualize the latent space of protein-ligand interactions encoded by IGModel and conduct interpretability analysis, providing valuable insights. This study presents a novel framework for DL-based prediction of protein-ligand interactions, contributing to the advancement of this field. The IGModel is available at GitHub repository https://github.com/zchwang/IGModel.

Machine Learning for Sequence and Structure-Based Protein-Ligand Interaction Prediction.

Developing new drugs is too expensive and time -consuming. Accurately predicting the interaction between drugs and targets will likely change how the drug is discovered. Machine learning-based protein-ligand interaction prediction has demonstrated significant potential. In this paper, computational methods, focusing on sequence and structure to study protein-ligand interactions, are examined. Therefore, this paper starts by presenting an overview of the data sets applied in this area, as well as the various approaches applied for representing proteins and ligands. Then, sequence-based and structure-based classification criteria are subsequently utilized to categorize and summarize both the classical machine learning models and deep learning models employed in protein-ligand interaction studies. Moreover, the evaluation methods and interpretability of these models are proposed. Furthermore, delving into the diverse applications of protein-ligand interaction models in drug research is presented. Lastly, the current challenges and future directions in this field are addressed.

Toward generalizable structure‐based deep learning models for protein–ligand interaction prediction: Challenges and strategies

AbstractAccurate and rapid prediction of protein–ligand interactions (PLIs) is the fundamental challenge of drug discovery. Deep learning methods have been harnessed for this purpose, yet the insufficient generalizability of PLI prediction prevents their broader impact on practical applications. Here, we highlight the significance of PLI model generalizability by conceiving PLIs as a function defined on infinitely diverse protein–ligand pairs and binding poses. To delve into the generalization challenges within PLI predictions, we comprehensively explore the evaluation strategies to assess the generalizability fairly. Moreover, we categorize structure‐based PLI models with leveraged strategies for learning generalizable features from structure‐based PLI data. Finally, we conclude the review by emphasizing the need for accurate pose‐predicting methods, which is a prerequisite for more accurate PLI predictions.This article is categorized under: Data Science > Artificial Intelligence/Machine Learning Data Science > Chemoinformatics Structure and Mechanism > Computational Biochemistry and Biophysics

PIGNet2: a versatile deep learning-based protein–ligand interaction prediction model for binding affinity scoring and virtual screening

PIGNet2, a versatile protein–ligand interaction prediction model that performs well in both molecule identification and optimization, demonstrates its potential in early-stage drug discovery.

Does protein pretrained language model facilitate the prediction of protein-ligand interaction?

Protein-ligand interaction (PLI) is a critical step for drug discovery. Recently, protein pretrained language models (PLMs) have showcased exceptional performance across a wide range of protein-related tasks. However, a significant heterogeneity exists between the PLM and PLI tasks, leading to a degree of uncertainty. In this study, we propose a method that quantitatively assesses the significance of protein PLMs in PLI prediction. Specifically, we analyze the performance of three widely-used protein PLMs (TAPE, ESM-1b, and ProtTrans) on three PLI tasks (PDBbind, Kinase, and DUD-E). The model with pre-training consistently achieves improved performance and decreased time cost, demonstrating that enhance both the accuracy and efficiency of PLI prediction. By quantitatively assessing the transferability, the optimal PLM for each PLI task is identified without the need for costly transfer experiments. Additionally, we examine the contributions of PLMs on the distribution of feature space, highlighting the improved discriminability after pre-training. Our findings provide insights into the mechanisms underlying PLMs in PLI prediction and pave the way for the design of more interpretable and accurate PLMs in the future. Code and data are freely available at https://github.com/brian-zZZ/PLM-PLI.

PocketAnchor: Learning structure-based pocket representations for protein-ligand interaction prediction.

Protein-ligand interactions are essential for cellular activities and drug discovery processes. Appropriately and effectively representing protein features is of vital importance for developing computational approaches, especially data-driven methods, for predicting protein-ligand interactions. However, existing approaches may not fully investigate the features of the ligand-occupying regions in the protein pockets. Here, we design a structure-based protein representation method, named PocketAnchor, for capturing the local environmental and spatial features of protein pockets to facilitate protein-ligand interaction-related learning tasks. We define "anchors" as probe points reaching into the cavities and those located near the surface of proteins, and we design a specific message passing strategy for gathering local information from the atoms and surface neighboring these anchors. Comprehensive evaluation of our method demonstrated its successful applications in pocket detection and binding affinity prediction, which indicated that our anchor-based approach can provide effective protein feature representations for improving the prediction of protein-ligand interactions.

AQDnet: Deep Neural Network for Protein-Ligand Docking Simulation.

We have developed an innovative system, AI QM Docking Net (AQDnet), which utilizes the three-dimensional structure of protein-ligand complexes to predict binding affinity. This system is novel in two respects: first, it significantly expands the training dataset by generating thousands of diverse ligand configurations for each protein-ligand complex and subsequently determining the binding energy of each configuration through quantum computation. Second, we have devised a method that incorporates the atom-centered symmetry function (ACSF), highly effective in describing molecular energies, for the prediction of protein-ligand interactions. These advancements have enabled us to effectively train a neural network to learn the protein-ligand quantum energy landscape (P-L QEL). Consequently, we have achieved a 92.6% top 1 success rate in the CASF-2016 docking power, placing first among all models assessed in the CASF-2016, thus demonstrating the exceptional docking performance of our model.

DeepBindGCN: Integrating Molecular Vector Representation with Graph Convolutional Neural Networks for Protein-Ligand Interaction Prediction.

The core of large-scale drug virtual screening is to select the binders accurately and efficiently with high affinity from large libraries of small molecules in which non-binders are usually dominant. The binding affinity is significantly influenced by the protein pocket, ligand spatial information, and residue types/atom types. Here, we used the pocket residues or ligand atoms as the nodes and constructed edges with the neighboring information to comprehensively represent the protein pocket or ligand information. Moreover, the model with pre-trained molecular vectors performed better than the one-hot representation. The main advantage of DeepBindGCN is that it is independent of docking conformation, and concisely keeps the spatial information and physical-chemical features. Using TIPE3 and PD-L1 dimer as proof-of-concept examples, we proposed a screening pipeline integrating DeepBindGCN and other methods to identify strong-binding-affinity compounds. It is the first time a non-complex-dependent model has achieved a root mean square error (RMSE) value of 1.4190 and Pearson r value of 0.7584 in the PDBbind v.2016 core set, respectively, thereby showing a comparable prediction power with the state-of-the-art affinity prediction models that rely upon the 3D complex. DeepBindGCN provides a powerful tool to predict the protein-ligand interaction and can be used in many important large-scale virtual screening application scenarios.

519. DeepLPI: A Novel Drug Repurposing Model based on Ligand-Protein Interaction Using Deep Learning

Abstract Background Drug repurposing has gained increased attention because it proposes to find effective cures for new diseases from approved drugs to lower development costs and time, and computational prediction of protein-ligand interactions (PLI) can provide accurate and fast drug screening. The shortcomings of existing machine learning-based methods for PLI prediction include 1) using human-selected features leads to loss of information and therefore lower accuracy and 2) using limited 3D structure data for input leads to lower generalizability. Methods To address the shortcomings, I proposed DeepLPI, a novel deep learning-based model that takes as input the raw sequences of drug molecules and proteins. DeepLPI applied pre-trained embedding models to encode the raw sequences into dense vector representations, which were then fed into 1D-CNN and biLSTM to obtain predictions. BindingDB dataset was used for model training and performance evaluation that is compared with a start-of-the-art method DeepCDA. Fig 1.Principle of NLP-inspired deep learning-based methods to embed drug SMILES strings (Mol2vec) and protein sequences (ProSE). Similar drugs or proteins will be in close distance after the embedding because these methods can capture contextual information.Fig 2.Model architecture from input to output of DeepLPI. Results Results showed that the DeepLPI reached an overall AUROC of 0.79 based on the BindingDB internal test, 76% better than DeepCDA. DeepLPI also outperformed DeepCDA using the external Davis dataset (AUROC=0.53) and COVID-19 3CL Protease dataset (AUROC=0.61). Figure 3.A randomly reserved part of the BindingDB dataset was grouped into four cases for testing, according to whether the drug molecules or the proteins are seen in the training set. (Note that the drug-target pairs in the test set are excluded from the training set.)Fig 4.We compare with DeepCDA, a recently published DTI prediction model that reported to be better than previous models. DeepLPI is 76% better in accuracy metric AUROC in classification tasks.Fig 5.As a preliminary attempt, the model trained on BindingDB was applied to a recent COVID-19 dataset targeting the 3CL-protease, which reported 897 small molecule drugs. Our model is better than the baseline method by 25%. Conclusion The high performance of DeepLPI suggests that our model has the potential to identify new COVID-19 drugs when applied to approved drugs. The generalizability of the model also promises applications to diseases in a wider scope. Disclosures All Authors: No reported disclosures.

DeepLPI: a novel deep learning-based model for protein–ligand interaction prediction for drug repurposing

The substantial cost of new drug research and development has consistently posed a huge burden for both pharmaceutical companies and patients. In order to lower the expenditure and development failure rate, repurposing existing and approved drugs by identifying interactions between drug molecules and target proteins based on computational methods have gained growing attention. Here, we propose the DeepLPI, a novel deep learning-based model that mainly consists of ResNet-based 1-dimensional convolutional neural network (1D CNN) and bi-directional long short term memory network (biLSTM), to establish an end-to-end framework for protein–ligand interaction prediction. We first encode the raw drug molecular sequences and target protein sequences into dense vector representations, which go through two ResNet-based 1D CNN modules to derive features, respectively. The extracted feature vectors are concatenated and further fed into the biLSTM network, followed by the MLP module to finally predict protein–ligand interaction. We downloaded the well-known BindingDB and Davis dataset for training and testing our DeepLPI model. We also applied DeepLPI on a COVID-19 dataset for externally evaluating the prediction ability of DeepLPI. To benchmark our model, we compared our DeepLPI with the baseline methods of DeepCDA and DeepDTA, and observed that our DeepLPI outperformed these methods, suggesting the high accuracy of the DeepLPI towards protein–ligand interaction prediction. The high prediction performance of DeepLPI on the different datasets displayed its high capability of protein–ligand interaction in generalization, demonstrating that the DeepLPI has the potential to pinpoint new drug-target interactions and to find better destinations for proven drugs.

P04-18 iSafeRat® SESAME-3D: prediction of protein-ligand interactions at the molecular level to detect endocrine disrupting potential of chemicals

P04-18 iSafeRat® SESAME-3D: prediction of protein-ligand interactions at the molecular level to detect endocrine disrupting potential of chemicals

PocketAnchor: Learning Structure-Based Pocket Representations for Protein-Ligand Interaction Prediction

PocketAnchor: Learning Structure-Based Pocket Representations for Protein-Ligand Interaction Prediction

Development of a graph convolutional neural network model for efficient prediction of protein-ligand binding affinities.

Prediction of protein-ligand interactions is a critical step during the initial phase of drug discovery. We propose a novel deep-learning-based prediction model based on a graph convolutional neural network, named GraphBAR, for protein-ligand binding affinity. Graph convolutional neural networks reduce the computational time and resources that are normally required by the traditional convolutional neural network models. In this technique, the structure of a protein-ligand complex is represented as a graph of multiple adjacency matrices whose entries are affected by distances, and a feature matrix that describes the molecular properties of the atoms. We evaluated the predictive power of GraphBAR for protein-ligand binding affinities by using PDBbind datasets and proved the efficiency of the graph convolution. Given the computational efficiency of graph convolutional neural networks, we also performed data augmentation to improve the model performance. We found that data augmentation with docking simulation data could improve the prediction accuracy although the improvement seems not to be significant. The high prediction performance and speed of GraphBAR suggest that such networks can serve as valuable tools in drug discovery.

Elucidating the multiple roles of hydration for accurate protein-ligand binding prediction via deep learning

Accurate and efficient prediction of protein-ligand interactions has been a long-lasting dream of practitioners in drug discovery. The insufficient treatment of hydration is widely recognized to be a major limitation for accurate protein-ligand scoring. Using an integration of molecular dynamics simulations on thousands of protein structures with novel big-data analytics based on convolutional neural networks and deep Taylor decomposition, we consistently identify here three different patterns of hydration to be essential for protein-ligand interactions. In addition to desolvation and water-mediated interactions, the formation of enthalpically favorable networks of first-shell water molecules around solvent-exposed ligand moieties is identified to be essential for protein-ligand binding. Despite being currently neglected in drug discovery, this hydration phenomenon could lead to new avenues in optimizing the free energy of ligand binding. Application of deep neural networks incorporating hydration to docking provides 89% accuracy in binding pose ranking, an essential step for rational structure-based drug design.

IN SILICO PREDICTION OF PROTEIN-LIGAND INTERACTIONS

The method for computer prediction of protein-ligand interactions was developed. The amino acid sequences of target proteins and structural descriptions of small molecule ligands are used as the input data. The method was tested on protein families representing perspective drug targets. The developed approach allows one to predict ligand-protein interactions with high efficiency.

A molecular interaction field describing nonconventional intermolecular interactions and its application to protein–ligand interaction prediction

Nonconventional noncovalent interactions such as CH/O, CH/π, and halogen bonds play important roles in molecular recognition in biological systems and have been increasingly exploited in structure-based drug design. In silico approaches that consider these interactions would be an effective strategy for drug discovery. The computation of the molecular interaction field (MIF), which is a three-dimensional (3D) potential map describing the interactions formed around a target compound, would assist the design of molecules that bind to the target biomolecule via nonconventional interactions. In this study, we developed a novel MIF calculation method that describes nonconventional interactions. This method evaluates the MIF as the interaction energy between the target ligand molecule and probe molecule. To describe the nonconventional interactions, our method employs quantum chemical calculations with four types of probe molecules. The calculated MIFs for casein kinase 2 (CK2) inhibitors correctly identify the halogen bond, CH/π, and CH/O interactions formed in the CK2/inhibitor complexes. Additionally, we have developed a method for calculating the protein–ligand interaction energy (Eint) based on the MIF and a coarse-grained protein model. The calculated interaction energies for CK2 inhibitors correlate with the experimental log(Ki) values. Thus, MIF and Eint obtained by our method show promise as descriptors for protein-ligand interaction prediction by considering nonconventional noncovalent interactions.

Predicting binding poses and affinity ranking in D3R Grand Challenge using PL-PatchSurfer2.0.

Computational prediction of protein-ligand interactions is a useful approach that aids the drug discovery process. Two major tasks of computational approaches are to predict the docking pose of a compound in a known binding pocket and to rank compounds in a library according to their predicted binding affinities. There are many computational tools developed in the past decades both in academia and industry. To objectively assess the performance of existing tools, the community has held a blind assessment of computational predictions, the Drug Design Data Resource Grand Challenge. This round, Grand Challenge 4 (GC4), focused on two targets, protein beta-secretase 1 (BACE-1) and cathepsin S (CatS). We participated in GC4 in both BACE-1 and CatS challenges using our molecular surface-based virtual screening method, PL-PatchSurfer2.0. A unique feature of PL-PatchSurfer2.0 is that it uses the three-dimensional Zernike descriptor, a mathematical moment-based shape descriptor, to quantify local shape complementarity between a ligand and a receptor, which properly incorporates molecular flexibility and provides stable affinity assessment for a bound ligand-receptor complex. Since PL-PatchSurfer2.0 does not explicitly build a bound pose of a ligand, we used an external docking program, such as AutoDock Vina, to provide an ensemble of poses, which were then evaluated by PL-PatchSurfer2.0. Here, we provide an overview of our method and report the performance in GC4.

Predicting protein-ligand interactions based on bow-pharmacological space and Bayesian additive regression trees

Identifying potential protein-ligand interactions is central to the field of drug discovery as it facilitates the identification of potential novel drug leads, contributes to advancement from hits to leads, predicts potential off-target explanations for side effects of approved drugs or candidates, as well as de-orphans phenotypic hits. For the rapid identification of protein-ligand interactions, we here present a novel chemogenomics algorithm for the prediction of protein-ligand interactions using a new machine learning approach and novel class of descriptor. The algorithm applies Bayesian Additive Regression Trees (BART) on a newly proposed proteochemical space, termed the bow-pharmacological space. The space spans three distinctive sub-spaces that cover the protein space, the ligand space, and the interaction space. Thereby, the model extends the scope of classical target prediction or chemogenomic modelling that relies on one or two of these subspaces. Our model demonstrated excellent prediction power, reaching accuracies of up to 94.5–98.4% when evaluated on four human target datasets constituting enzymes, nuclear receptors, ion channels, and G-protein-coupled receptors . BART provided a reliable probabilistic description of the likelihood of interaction between proteins and ligands, which can be used in the prioritization of assays to be performed in both discovery and vigilance phases of small molecule development.

An Overview of Scoring Functions Used for Protein-Ligand Interactions in Molecular Docking.

Currently, molecular docking is becoming a key tool in drug discovery and molecular modeling applications. The reliability of molecular docking depends on the accuracy of the adopted scoring function, which can guide and determine the ligand poses when thousands of possible poses of ligand are generated. The scoring function can be used to determine the binding mode and site of a ligand, predict binding affinity and identify the potential drug leads for a given protein target. Despite intensive research over the years, accurate and rapid prediction of protein-ligand interactions is still a challenge in molecular docking. For this reason, this study reviews four basic types of scoring functions, physics-based, empirical, knowledge-based, and machine learning-based scoring functions, based on an up-to-date classification scheme. We not only discuss the foundations of the four types scoring functions, suitable application areas and shortcomings, but also discuss challenges and potential future study directions.

A Diverse Benchmark Based on 3D Matched Molecular Pairs for Validating Scoring Functions.

The prediction of protein–ligand interactions and their corresponding binding free energy is a challenging task in structure-based drug design and related applications. Docking and scoring is broadly used to propose the binding mode and underlying interactions as well as to provide a measure for ligand affinity or differentiate between active and inactive ligands. Various studies have revealed that most docking software packages reliably predict the binding mode, although scoring remains a challenge. Here, a diverse benchmark data set of 99 matched molecular pairs (3D-MMPs) with experimentally determined X-ray structures and corresponding binding affinities is introduced. This data set was used to study the predictive power of 13 commonly used scoring functions to demonstrate the applicability of the 3D-MMP data set as a valuable tool for benchmarking scoring functions.