Protein Binding Site Prediction Research Articles

Turbocharging protein binding site prediction with geometric attention, inter-resolution transfer learning, and homology-based augmentation

BackgroundLocating small molecule binding sites in target proteins, in the resolution of either pocket or residue, is critical in many drug-discovery scenarios. Since it is not always easy to find such binding sites using conventional methods, different deep learning methods to predict binding sites out of protein structures have been developed in recent years. The existing deep learning based methods have several limitations, including (1) the inefficiency of the CNN-only architecture, (2) loss of information due to excessive post-processing, and (3) the under-utilization of available data sources.MethodsWe present a new model architecture and training method that resolves the aforementioned problems. First, by layering geometric self-attention units on top of residue-level 3D CNN outputs, our model overcomes the problems of CNN-only architectures. Second, by configuring the fundamental units of computation as residues and pockets instead of voxels, our method reduced the information loss from post-processing. Lastly, by employing inter-resolution transfer learning and homology-based augmentation, our method maximizes the utilization of available data sources to a significant extent.ResultsThe proposed method significantly outperformed all state-of-the-art baselines regarding both resolutions—pocket and residue. An ablation study demonstrated the indispensability of our proposed architecture, as well as transfer learning and homology-based augmentation, for achieving optimal performance. We further scrutinized our model’s performance through a case study involving human serum albumin, which demonstrated our model’s superior capability in identifying multiple binding sites of the protein, outperforming the existing methods.ConclusionsWe believe that our contribution to the literature is twofold. Firstly, we introduce a novel computational method for binding site prediction with practical applications, substantiated by its strong performance across diverse benchmarks and case studies. Secondly, the innovative aspects in our method— specifically, the design of the model architecture, inter-resolution transfer learning, and homology-based augmentation—would serve as useful components for future work.

Advances in the Application of Protein Language Modeling for Nucleic Acid Protein Binding Site Prediction.

Protein and nucleic acid binding site prediction is a critical computational task that benefits a wide range of biological processes. Previous studies have shown that feature selection holds particular significance for this prediction task, making the generation of more discriminative features a key area of interest for many researchers. Recent progress has shown the power of protein language models in handling protein sequences, in leveraging the strengths of attention networks, and in successful applications to tasks such as protein structure prediction. This naturally raises the question of the applicability of protein language models in predicting protein and nucleic acid binding sites. Various approaches have explored this potential. This paper first describes the development of protein language models. Then, a systematic review of the latest methods for predicting protein and nucleic acid binding sites is conducted by covering benchmark sets, feature generation methods, performance comparisons, and feature ablation studies. These comparisons demonstrate the importance of protein language models for the prediction task. Finally, the paper discusses the challenges of protein and nucleic acid binding site prediction and proposes possible research directions and future trends. The purpose of this survey is to furnish researchers with actionable suggestions for comprehending the methodologies used in predicting protein-nucleic acid binding sites, fostering the creation of protein-centric language models, and tackling real-world obstacles encountered in this field.

Machine Learning in Enhancing Protein Binding Sites Predictions - What Has Changed Since Then?

Accurate identification of protein binding sites is pivotal for understanding molecular interactions and facilitating drug discovery efforts. However, the dynamic nature of proteinligand interactions presents a formidable challenge, necessitating innovative approaches to bridge the gap between theoretical predictions and experimental realities. This review explores the challenges and recent advancements in protein binding site prediction. Specifically, we highlight the integration of molecular dynamics simulations, machine learning, and deep learning techniques to capture the dynamic and complex nature of protein-ligand interactions. Additionally, we discuss the importance of integrating experimental data, such as structural information and biochemical assays, to enhance prediction accuracy and reliability. By navigating the intersection of classical and the onset of machine learning and deep learning approaches, we aim to provide insights into current state-of-the-art techniques and chart a course for future protein binding site prediction advancements. Ultimately, these efforts could unravel the mysteries of protein-ligand interactions and accelerate drug discovery endeavors.

DeepProSite: structure-aware protein binding site prediction using ESMFold and pretrained language model.

Identifying the functional sites of a protein, such as the binding sites of proteins, peptides, or other biological components, is crucial for understanding related biological processes and drug design. However, existing sequence-based methods have limited predictive accuracy, as they only consider sequence-adjacent contextual features and lack structural information. In this study, DeepProSite is presented as a new framework for identifying protein binding site that utilizes protein structure and sequence information. DeepProSite first generates protein structures from ESMFold and sequence representations from pretrained language models. It then uses Graph Transformer and formulates binding site predictions as graph node classifications. In predicting protein-protein/peptide binding sites, DeepProSite outperforms state-of-the-art sequence- and structure-based methods on most metrics. Moreover, DeepProSite maintains its performance when predicting unbound structures, in contrast to competing structure-based prediction methods. DeepProSite is also extended to the prediction of binding sites for nucleic acids and other ligands, verifying its generalization capability. Finally, an online server for predicting multiple types of residue is established as the implementation of the proposed DeepProSite. The datasets and source codes can be accessed at https://github.com/WeiLab-Biology/DeepProSite. The proposed DeepProSite can be accessed at https://inner.wei-group.net/DeepProSite/.

RefinePocket: An Attention-Enhanced and Mask-Guided Deep Learning Approach for Protein Binding Site Prediction.

Protein binding site prediction is an important prerequisite task of drug discovery and design. While binding sites are very small, irregular and varied in shape, making the prediction very challenging. Standard 3D U-Net has been adopted to predict binding sites but got stuck with unsatisfactory prediction results, incomplete, out-of-bounds, or even failed. The reason is that this scheme is less capable of extracting the chemical interactions of the entire region and hardly takes into account the difficulty of segmenting complex shapes. In this paper, we propose a refined U-Net architecture, called RefinePocket, consisting of an attention-enhanced encoder and a mask-guided decoder. During encoding, taking binding site proposal as input, we employ Dual Attention Block (DAB) hierarchically to capture rich global information, exploring residue relationship and chemical correlations in spatial and channel dimensions respectively. Then, based on the enhanced representation extracted by the encoder, we devise Refine Block (RB) in the decoder to enable self-guided refinement of uncertain regions gradually, resulting in more precise segmentation. Experiments show that DAB and RB complement and promote each other, making RefinePocket has an average improvement of 10.02% on DCC and 4.26% on DVO compared with the state-of-the-art method on four test sets.

Spatial Layouts of Low-Entropy Hydration Shells Guide Protein Binding.

Protein-protein binding enables orderly biological self-organization and is therefore considered a miracle of nature. Protein‒protein binding is driven by electrostatic forces, hydrogen bonding, van der Waals force, and hydrophobic interactions. Among these physical forces, only hydrophobic interactions can be considered long-range intermolecular attractions between proteins due to the electrostatic shielding of surrounding water molecules. Low-entropy hydration shells around proteins drive hydrophobic attraction among them that essentially coordinate protein‒protein binding. Here, an innovative method is developed for identifying low-entropy regions of hydration shells of proteins by screening off pseudohydrophilic groups on protein surfaces and revealing that large low-entropy regions of the hydration shells typically cover the binding sites of individual proteins. According to an analysis of determined protein complex structures, shape matching between a large low-entropy hydration shell region of a protein and that of its partner at the binding sites is revealed as a universal law. Protein‒protein binding is thus found to be mainly guided by hydrophobic collapse between the shape-matched low-entropy hydration shells that is verified by bioinformatics analyses of hundreds of structures of protein complexes, which cover four test systems. A simple algorithm is proposed to accurately predict protein binding sites.

CAT-Site: Predicting Protein Binding Sites Using a Convolutional Neural Network.

Identifying binding sites on the protein surface is an important part of computer-assisted drug design processes. Reliable prediction of binding sites not only assists with docking algorithms, but it can also explain the possible side-effects of a potential drug as well as its efficiency. In this work, we propose a novel workflow for predicting possible binding sites of a ligand on a protein surface. We use proteins from the PDBbind and sc-PDB databases, from which we combine available ligand information for similar proteins using all the possible ligands rather than only a special sub-selection to generalize the work of existing research. After performing protein clustering and merging of ligands of similar proteins, we use a three-dimensional convolutional neural network that takes into account the spatial structure of a protein. Lastly, we combine ligandability predictions for points on protein surfaces into joint binding sites. Analysis of our model's performance shows that its achieved sensitivity is 0.829, specificity is 0.98, and F1 score is 0.517, and that for 54% of larger and pharmacologically relevant binding sites, the distance between their real and predicted centers amounts to less than 4 Å.

Protein binding sites for drug design

Drug development is a lengthy and challenging process that can be accelerated at early stages by new mathematical approaches and modern computers. To address this important issue, we are developing new mathematical solutions for the detection and characterization of protein binding sites that are important for new drug development. In this review, we present algorithms based on graph theory combined with molecular dynamics simulations that we have developed for studying biological target proteins to provide important data for optimizing the early stages of new drug development. A particular focus is the development of new protein binding site prediction algorithms (ProBiS) and new web tools for modeling pharmaceutically interesting molecules-ProBiS Tools (algorithm, database, web server), which have evolved into a full-fledged graphical tool for studying proteins in the proteome. ProBiS differs from other structural algorithms in that it can align proteins with different folds without prior knowledge of the binding sites. It allows detection of similar binding sites and can predict molecular ligands of various types of pharmaceutical interest that could be advanced to drugs to treat a disease, based on the entire Protein Data Bank (PDB) and AlphaFold database, including proteins not yet in the PDB. All ProBiS Tools are freely available to the academic community at http://insilab.org and https://probis.nih.gov.

Evaluation of the Effectiveness of Derived Features of AlphaFold2 on Single-Sequence Protein Binding Site Prediction.

Simple SummaryWith the development of artificial intelligence, researchers can roughly predict the crystal structure of a protein by computer without the need for biological experiments, which provides new ideas and solutions to problems, such as protein-protein interaction and drug-target predictions. In this study, we proposed strategies to combine predicted protein structures with deep learning networks and evaluated them on different protein binding site prediction tasks. Our computational experiment results showed that all proposed strategies could effectively encode structural information for deep learning models.Though AlphaFold2 has attained considerably high precision on protein structure prediction, it is reported that directly inputting coordinates into deep learning networks cannot achieve desirable results on downstream tasks. Thus, how to process and encode the predicted results into effective forms that deep learning models can understand to improve the performance of downstream tasks is worth exploring. In this study, we tested the effects of five processing strategies of coordinates on two single-sequence protein binding site prediction tasks. These five strategies are spatial filtering, the singular value decomposition of a distance map, calculating the secondary structure feature, and the relative accessible surface area feature of proteins. The computational experiment results showed that all strategies were suitable and effective methods to encode structural information for deep learning models. In addition, by performing a case study of a mutated protein, we showed that the spatial filtering strategy could introduce structural changes into HHblits profiles and deep learning networks when protein mutation happens. In sum, this work provides new insight into the downstream tasks of protein-molecule interaction prediction, such as predicting the binding residues of proteins and estimating the effects of mutations.

Flexibility-aware graph model for accurate epitope identification

Structural flexibility is an intrinsic characteristics of a protein upon interacting with other molecules, which mainly comes from the movement of a residue’s side chain, backbone and even an entire domain. Considering this property can be very helpful in protein binding analysis, such as epitope identification during antibody–antigen interaction. Unfortunately, to our knowledge, no approach is available at studying the dynamicity of protein binding from the computational perspective. We are pioneering a new perspective of exploring protein binding sites with considering the structural flexibility, particularly from the epitopes identification angle in antibody–antigen binding. To this end, we first obtained protein antigen structures with epitopes available, and built residue-level graphs of antigens. These graphs were highly densified subsequently by incorporating the structural flexibility. Later, the edge enriched graphs were clustered into overlapping subgraphs and were classified as epitope or non-epitope by a graph convolutional network. Experiments on epitope identification shown that the proposed flexibility-aware model markedly outperformed existing approaches by lifting the F1-score to 0.656, making a remarkable increment of 16.3% compared to the state-of-the-art. A quick study on generic protein binding site prediction also made a noteworthy improvement with increasing the F1-score by 8%. The superior performance obtained from both the specific and generic protein interaction analysis demonstrate that incorporating flexibility in computational models is helpful to strength the capability of identifying epitopes as well as general protein binding sites. This seminal study can be inspiring and promising to the wide range of protein interaction analysis.

ProB-Site: Protein Binding Site Prediction Using Local Features.

Protein–protein interactions (PPIs) are responsible for various essential biological processes. This information can help develop a new drug against diseases. Various experimental methods have been employed for this purpose; however, their application is limited by their cost and time consumption. Alternatively, computational methods are considered viable means to achieve this crucial task. Various techniques have been explored in the literature using the sequential information of amino acids in a protein sequence, including machine learning and deep learning techniques. The current efficiency of interaction-site prediction still has growth potential. Hence, a deep neural network-based model, ProB-site, is proposed. ProB-site utilizes sequential information of a protein to predict its binding sites. The proposed model uses evolutionary information and predicted structural information extracted from sequential information of proteins, generating three unique feature sets for every amino acid in a protein sequence. Then, these feature sets are fed to their respective sub-CNN architecture to acquire complex features. Finally, the acquired features are concatenated and classified using fully connected layers. This methodology performed better than state-of-the-art techniques because of the selection of the best features and contemplation of local information of each amino acid.

ScanNet: an interpretable geometric deep learning model for structure-based protein binding site prediction.

Predicting the functional sites of a protein from its structure, such as the binding sites of small molecules, other proteins or antibodies, sheds light on its function in vivo. Currently, two classes of methods prevail: machine learning models built on top of handcrafted features and comparative modeling. They are, respectively, limited by the expressivity of the handcrafted features and the availability of similar proteins. Here, we introduce ScanNet, an end-to-end, interpretable geometric deep learning model that learns features directly from 3D structures. ScanNet builds representations of atoms and amino acids based on the spatio-chemical arrangement of their neighbors. We train ScanNet for detecting protein-protein and protein-antibody binding sites, demonstrate its accuracy-including for unseen protein folds-and interpret the filters learned. Finally, we predict epitopes of the SARS-CoV-2 spike protein, validating known antigenic regions and predicting previously uncharacterized ones. Overall, ScanNet is a versatile, powerful and interpretable model suitable for functional site prediction tasks. A webserver for ScanNet is available from http://bioinfo3d.cs.tau.ac.il/ScanNet/ .

DFpin: Deep learning–based protein-binding site prediction with feature-based non-redundancy from RNA level

The interaction between proteins and RNA is closely related to various human diseases. Computer-aided drug design can be facilitated by detecting the RNA sites that bind proteins. However, due to the aggregation of binding sites in RNA sequences, high sample similarity occurs when extracting RNA fragments by using a sliding window. Considering these problems, we present a method, DFpin, to predict protein-interacting nucleotides in RNA. To retain more key nucleotide sites, we used the redundancy method based on feature similarity, that is, feature redundancy is removed based on the RNA mono-nucleotide composition to maintain the diversity of RNA samples and avoid the residue of redundant data. In addition, to extract key abstract features and avoid over-fitting, we used the cascade structure of a deep forest model to predict protein-interacting nucleotides. Overall, DFpin demonstrated excellent classification with 85.4% accuracy and 93.3% area under the curve. Compared with other methods, the accuracy of DFpin was better, suggesting that feature-based redundancy removal and deep forest can help predict nucleotides of protein interactions. The source code and all dataset are available at: https://github.com/zhaoxj-tech/DFpin.git.

SNB‐PSSM: A spatial neighbor‐based PSSM used for protein–RNA binding site prediction

Protein-RNA interactions play essential roles in a wide variety of biological processes. Recognition of RNA-binding residues on proteins has been a challenging problem. Most of methods utilize the position-specific scoring matrix (PSSM). It has been found that considering the evolutionary information of sequence neighboring residues can improve the prediction. In this work, we introduce a novel method SNB-PSSM (spatial neighbor-based PSSM) combined with the structure window scheme where the evolutionary information of spatially neighboring residues is considered. The results show our method consistently outperforms the standard and smoothed PSSM methods. Tested on multiple datasets, this approach shows an encouraging performance compared with RNABindRPlus, BindN+, PPRInt, xypan, Predict_RBP, SpaPF, PRNA, and KYG, although is inferior to RNAProSite, RBscore, and aaRNA. In addition, since our method is not sensitive to protein structure changes, it can be applied well on binding site predictions of modeled structures. Thus, the result also suggests the evolution of binding sites is spatially cooperative. The proposed method as an effective tool of considering evolutionary information can be widely used for the nucleic acid-/protein-binding site prediction and functional motif finding.

A Review About RNA–Protein-Binding Sites Prediction Based on Deep Learning

RNA-binding proteins (RBPs) play crucial roles in gene regulation. The advent of high-throughput experimental methods, has generated a huge volume of experimentally verified binding sites of RNA-binding proteins and greatly advanced the genome-wide studies of RNA-protein interactions. Many computational approaches have been proposed, including deep learning models, which have achieved remarkable performance on the identification of RNA-protein binding affinities and sites. In this review, we discuss machine learning and deep learning approaches, mainly focusing on the prediction of RNA and proteins binding sites on RNAs by deep learning. Furthermore, we discuss the advantages and disadvantages of these approaches. The workflow of deep learning is also revealed. We recommend some promising future directions of deep learning models in the study of RBP-binding sites on RNAs, especially the embedding, generative adversarial net, and attention model. Extraction and visualization methods involving motif are illustrated. Finally, we summarize the previous studies, and then compare the performance on different dataset.

Computational methods and tools for binding site recognition between proteins and small molecules: from classical geometrical approaches to modern machine learning strategies.

In the current "genomic era" the number of identified genes is growing exponentially. However, the biological function of a large number of the corresponding proteins is still unknown. Recognition of small molecule ligands (e.g., substrates, inhibitors, allosteric regulators, etc.) is pivotal for protein functions in the vast majority of the cases and knowledge of the region where these processes take place is essential for protein function prediction and drug design. In this regard, computational methods represent essential tools to tackle this problem. A significant number of software tools have been developed in the last few years which exploit either protein sequence information, structure information or both. This review describes the most recent developments in protein function recognition and binding site prediction, in terms of both freely-available and commercial solutions and tools, detailing the main characteristics of the considered tools and providing a comparative analysis of their performance.

FRSite: Protein drug binding site prediction based on faster R–CNN

Prediction of binding affinity of proteins and small molecules is a key step in drug design, and the location of binding sites is crucial for affinity prediction and molecular docking. In order to improve the accuracy of binding site prediction, a method called FRSite which improves the Faster R–CNN for protein binding site prediction is proposed in this paper. Multi-channel descriptors for proteins are generated to three dimensional (3D) girds and fed into the proposed Region Proposal Network (RPN-3D) network for potential proposals detection. Moreover, a 3D classifier is used to predict the bounding box of the binding site for a protein, and could also predict the center and size of the site. It can be seen from our comparative experiments that the proposed method can assist drug design.

Molecular characterization of the viral structural gene of the first dengue virus type 1 outbreak in Xishuangbanna: A border area of China, Burma and Laos

BackgroundXishuangbanna, a border area of China, Burma and Laos, had its first major DENV-1 outbreak in 2017. This study aims to explore the genetic characterization, potential source and evolution of the viruses in outbreak. MethodsThe structural protein C/prM/E genes of viruses isolated from local residents or Burmese travelers were sequenced followed by mutation, phylogenetic, homologous recombination, molecular clock and demographic reconstruction analysis. ResultsPhylogenetic analysis revealed that all of the strains were classified as three cluster of DENV-1. Cluster 1, 2 and 3 were most similar to China Guangzhou 2011, China Hubei 2014 and Laos 2008 strain, respectively. Among 236 base mutations, 31 caused nonsynonymous mutations when compared with the DENV-1SS. No homologous recombination signal was discovered. The structural protein of these strains had similar three-dimensional structure. Only site 434 showed differences among five predicted protein binding sites. Molecular clock phylogenetic and demographic reconstruction analysis showed that DENV-1 became highly diversified in 1972 followed by a slightly decreased period until 2017. ConclusionsDengue isolated strains show diversification between Burma and China. Amino acid substitution (I440T) may lead to weakened virulence of the epidemic strains. DENV-1 became highly diversified in 1972 followed by a slightly decreased period.

Evolutionary couplings and sequence variation effect predict protein binding sites.

Binding small ligands such as ions or macromolecules such as DNA, RNA, and other proteins is one important aspect of the molecular function of proteins. Many binding sites remain without experimental annotations. Predicting binding sites on a per-residue level is challenging, but if 3D structures are known, information about coevolving residue pairs (evolutionary couplings) can predict catalytic residues through mutual information. Here, we predicted protein binding sites from evolutionary couplings derived from a global statistical model using maximum entropy. Additionally, we included information from sequence variation. A simple method using a weighted sum over eight scores substantially outperformed random (F1 = 19.3% ± 0.7% vs F1 = 2% for random). Training a neural network on these eight scores (along with predicted solvent accessibility and conservation in protein families) improved substantially (F1 = 26.2% ±0.8%). Although the machine learning was limited by the small data set and possibly wrong annotations of binding sites, the predicted binding sites formed spatial clusters in the protein. The source code of the binding site predictions is available through GitHub: https://github.com/Rostlab/bindPredict.

Geometric and amino acid type determinants for protein‐protein interaction interfaces

BackgroundProtein‐protein interactions are essential to many biological processes. The binding site information of protein‐protein complexes is extremely useful to obtain their structures from biochemical experiments. Geometric description of protein structures is the precondition of protein binding site prediction and protein‐protein interaction analysis. The previous description of protein surface residues is incomplete, and little attention are paid to the implication of residue types for binding site prediction.MethodsHere, we found three new geometric features to characterize protein surface residues which are very effective for protein‐protein interface residue prediction. The new features and several commonly used descriptors were employed to train millions of residue type‐nonspecific or specific protein binding site predictors.ResultsThe amino acid type‐specific predictors are superior to the models without distinction of amino acid types. The performances of the best predictors are much better than those of the sophisticated methods developed before.ConclusionsThe results demonstrate that the geometric properties and amino acid types are very likely to determine if a protein surface residue would become an interface one when the protein binds to its partner.