G-LoSA: An efficient computational tool for local structure-centric biological studies and drug design

VirtuousPocketome: a computational tool for screening protein-ligand complexes to identify similar binding sites

Protein residues within binding pockets play a critical role in determining the range of ligands that can interact with a protein, influencing its structure and function. Identifying structural similarities in proteins offers valuable insights into their function and activation mechanisms, aiding in predicting protein-ligand interactions, anticipating off-target effects, and facilitating the development of therapeutic agents. Numerous computational methods assessing global or local similarity in protein cavities have emerged, but their utilization is impeded by complexity, impractical automation for amino acid pattern searches, and an inability to evaluate the dynamics of scrutinized protein-ligand systems. Here, we present a general, automatic and unbiased computational pipeline, named VirtuousPocketome, aimed at screening huge databases of proteins for similar binding pockets starting from an interested protein-ligand complex. We demonstrate the pipeline's potential by exploring a recently-solved human bitter taste receptor, i.e. the TAS2R46, complexed with strychnine. We pinpointed 145 proteins sharing similar binding sites compared to the analysed bitter taste receptor and the enrichment analysis highlighted the related biological processes, molecular functions and cellular components. This work represents the foundation for future studies aimed at understanding the effective role of tastants outside the gustatory system: this could pave the way towards the rationalization of the diet as a supplement to standard pharmacological treatments and the design of novel tastants-inspired compounds to target other proteins involved in specific diseases or disorders. The proposed pipeline is publicly accessible, can be applied to any protein-ligand complex, and could be expanded to screen any database of protein structures.

Pharmacophore-guided drug design using LdNMT as a model drug target for leishmaniasis

Leishmaniasis is caused by Leishmania genus parasites and has a high mortality rate. The available drugs to treat leishmaniasis fail due to acquired resistance in parasites. Several enzymes of the Leishmania parasite have been used to design new therapeutic molecules against leishmaniasis. This study uses a pharmacophore-guided approach to design the drug candidate by targeting Leishmania N-Myristoyl transferase (LdNMT). From the initial sequence analysis of LdNMT, we have identified a unique 20 amino acid stretch exploited for screening and designing the small molecules. The pharmacophore for the myristate binding site on LdNMT was elucidated, and a heatmap was constructed. The leishmanial NMT pharmacophore has similarities with other pathogenic microorganisms. Moreover, substituting alanine in pharmacophoric residues elevates the affinity of myristate with NMT. Furthermore, a molecular dynamics (MD) simulation study was conducted to ascertain the stability of the mutants and or wild type. The wild-type NMT has a comparatively low affinity to myristate compared to alanine mutants, indicating that hydrophobic residues favor the myristate binding. The molecules were initially designed by using pharmacophore as a sieving mechanism. In subsequent steps, the selected molecules screened against leishmanial unique amino acid stretch and subsequently with human, leishmanial full-size NMTs. The compounds BP5, TYI, DMU, 3PE and 4UL were the top hits and chemical features similar to the myristate. The molecule 4UL was found to be highly specific towards leishmanial NMT over human NMT, suggesting the molecule is a strong leishmanial NMT inhibitor. The molecule can be taken further to assess it in in-vitro conditions.

Structural similarities between SAM and ATP recognition motifs and detection of ATP binding in a SAM binding DNA methyltransferase

S-adenosylmethionine (SAM) is a ubiquitous co-factor that serves as a donor for methylation reactions and additionally serves as a donor of other functional groups such as amino and ribosyl moieties in a variety of other biochemical reactions. Such versatility in function is enabled by the ability of SAM to be recognized by a wide variety of protein molecules that vary in their sequences and structural folds. To understand what gives rise to specific SAM binding in diverse proteins, we set out to study if there are any structural patterns at their binding sites. A comprehensive analysis of structures of the binding sites of SAM by all-pair comparison and clustering, indicated the presence of 4 different site-types, only one among them being well studied. For each site-type we decipher the common minimum principle involved in SAM recognition by diverse proteins and derive structural motifs that are characteristic of SAM binding. The presence of the structural motifs with precise three-dimensional arrangement of amino acids in SAM sites that appear to have evolved independently, indicates that these are winning arrangements of residues to bring about SAM recognition. Further, we find high similarity between one of the SAM site types and a well known ATP binding site type. We demonstrate using in vitro experiments that a known SAM binding protein, HpyAII.M1, a type 2 methyltransferase can bind and hydrolyse ATP. We find common structural motifs that explain this, further supported through site-directed mutagenesis. Observation of similar motifs for binding two of the most ubiquitous ligands in multiple protein families with diverse sequences and structural folds presents compelling evidence at the molecular level in favour of convergent evolution.

CHARMM-GUI-Based Induced Fit Docking Workflow to Generate Reliable Protein-Ligand Binding Modes

Molecular docking is a preferred method to predict ligand binding modes and their binding energy to target protein receptors, which is critical in early phase structure-based drug discovery. However, there is a persistent challenge in docking that can be attributed to the induced fit effect, as receptor binding sites undergo induced fit conformational changes upon ligand binding to achieve better binding modes. In this work, based on CHARMM-GUI LBS Finder& Refiner and High-Throughput Simulator, we present a straightforward CHARMM-GUI induced fit docking (CGUI-IFD) workflow to generate reliable protein-ligand binding modes. The CGUI-IFD workflow generates an ensemble of receptor binding site conformations through ligand-binding site (LBS) refinement, runs rigid receptor docking, and performs high-throughput molecular dynamics (MD) simulations of protein-ligand complex structures in explicit solvents. The results are evaluated based on the ligand root-mean-square deviation (RMSD)-based binding stability and the molecular mechanics generalized Born surface area binding energy. For a benchmark test, we used 258 cross-docking protein-ligand pairs across 41 target proteins from the Schrodinger IFD-MD data set. The application of CGUI-IFD on this data set shows 80% success rate (within 2.5 Å RMSD from the experimental structures). We expect that the CGUI-IFD workflow can be useful to generate reliable ligand binding modes for cross-docking cases.

HLA-Clus: HLA class I clustering based on 3D structure

BACKGROUND

In a previous paper, we classified populated HLA class I alleles into supertypes and subtypes based on the similarity of 3D landscape of peptide binding grooves, using newly defined structure distance metric and hierarchical clustering approach. Compared to other approaches, our method achieves higher correlation with peptide binding specificity, intra-cluster similarity (cohesion), and robustness. Here we introduce HLA-Clus, a Python package for clustering HLA Class I alleles using the method we developed recently and describe additional features including a new nearest neighbor clustering method that facilitates clustering based on user-defined criteria.

RESULTS

The HLA-Clus pipeline includes three stages: First, HLA Class I structural models are coarse grained and transformed into clouds of labeled points. Second, similarities between alleles are determined using a newly defined structure distance metric that accounts for spatial and physicochemical similarities. Finally, alleles are clustered via hierarchical or nearest-neighbor approaches. We also interfaced HLA-Clus with the peptide:HLA affinity predictor MHCnuggets. By using the nearest neighbor clustering method to select optimal allele-specific deep learning models in MHCnuggets, the average accuracy of peptide binding prediction of rare alleles was improved.

CONCLUSIONS

The HLA-Clus package offers a solution for characterizing the peptide binding specificities of a large number of HLA alleles. This method can be applied in HLA functional studies, such as the development of peptide affinity predictors, disease association studies, and HLA matching for grafting. HLA-Clus is freely available at our GitHub repository ( https://github.com/yshen25/HLA-Clus ).

Identification of Fungal and Bacterial Denitrification Inhibitors Targeting Copper Nitrite Reductase

The usage of nitrification inhibitors is one of the strategies that reduce or slow down the denitrification process to prevent nitrogen loss to the atmosphere in the form of N₂O. Directly targeting microbial denitrification could be one of the mitigation strategies; however, until now little efforts have been devoted toward the development of denitrification inhibitors. Here, we have identified small-molecule inhibitors of one of the proteins involved in the fungal denitrification pathway. Specifically, virtual screening was employed to identify the inhibitors of copper-containing nitrite reductase (FoNirK) of the filamentous fungus Fusarium oxysporum. Three series of chemical compounds were identified, out of which compounds belonging to two chemical scaffolds inhibited FoNirK enzymatic activity in low micromolar ranges. Several compounds also displayed moderate inhibition of fungal denitrification activity in vivo. Evaluation of in vitro activity against NirK from denitrifying bacterium Achromobacter xylosoxidans (AxNirK) and in vivo bacterial denitrification revealed a similar inhibitory profile.

HLA Class I Supertype Classification Based on Structural Similarity

HLA class I proteins, a critical component in adaptive immunity, bind and present intracellular Ags to CD8+ T cells. The extreme polymorphism of HLA genes and associated peptide binding specificities leads to challenges in various endeavors, including neoantigen vaccine development, disease association studies, and HLA typing. Supertype classification, defined by clustering functionally similar HLA alleles, has proven helpful in reducing the complexity of distinguishing alleles. However, determining supertypes via experiments is impractical, and current in silico classification methods exhibit limitations in stability and functional relevance. In this study, by incorporating three-dimensional structures we present a method for classifying HLA class I molecules with improved breadth, accuracy, stability, and flexibility. Critical for these advances is our finding that structural similarity highly correlates with peptide binding specificity. The new classification should be broadly useful in peptide-based vaccine development and HLA-disease association studies.

Holo Protein Conformation Generation from Apo Structures by Ligand Binding Site Refinement

An important part in structure-based drug design is the selection of an appropriate protein structure. It has been revealed that a holo protein structure that contains a well-defined binding site is a much better choice than an apo structure in structure-based drug discovery. Therefore, it is valuable to obtain a holo-like protein conformation from apo structures in the case where no holo structure is available. Meeting the need, we present a robust approach to generate reliable holo-like structures from apo structures by ligand binding site refinement with restraints derived from holo templates with low homology. Our method was tested on a test set of 32 proteins from the DUD-E data set and compared with other approaches. It was shown that our method successfully refined the apo structures toward the corresponding holo conformations for 23 of 32 proteins, reducing the average all-heavy-atom RMSD of binding site residues by 0.48 Å. In addition, when evaluated against all the holo structures in the protein data bank, our method can improve the binding site RMSD for 14 of 19 cases that experience significant conformational changes. Furthermore, our refined structures also demonstrate their advantages over the apo structures in ligand binding mode predictions by both rigid docking and flexible docking and in virtual screening on the database of active and decoy ligands from the DUD-E. These results indicate that our method is effective in recovering holo-like conformations and will be valuable in structure-based drug discovery.

Estimating the Similarity between Protein Pockets

With the exponential increase in publicly available protein structures, the comparison of protein binding sites naturally emerged as a scientific topic to explain observations or generate hypotheses for ligand design, notably to predict ligand selectivity for on- and off-targets, explain polypharmacology, and design target-focused libraries. The current review summarizes the state-of-the-art computational methods applied to pocket detection and comparison as well as structural druggability estimates. The major strengths and weaknesses of current pocket descriptors, alignment methods, and similarity search algorithms are presented. Lastly, an exhaustive survey of both retrospective and prospective applications in diverse medicinal chemistry scenarios illustrates the capability of the existing methods and the hurdle that still needs to be overcome for more accurate predictions.

Fast Local Alignment of Protein Pockets (FLAPP): A System-Compiled Program for Large-Scale Binding Site Alignment

Protein function is a direct consequence of its sequence, structure, and the arrangement at the binding site. Bioinformatics using sequence analysis is typically used to gain a first insight into protein function. Protein structures, on the other hand, provide a higher resolution platform into understanding functions. As the protein structural information is increasingly becoming available through experimental structure determination and through advances in computational methods for structure prediction, the opportunity to utilize these data is also increasing. Structural analysis of small molecule ligand binding sites in particular provides a direct and more accurate window to infer protein function. However, it remains a poorly utilized resource due to the huge computational cost of existing methods that make large-scale structural comparisons of binding sites prohibitive. Here, we present an algorithm called FLAPP that produces very rapid atomic level alignments. By combining clique matching in graphs and the power of modern CPU architectures, FLAPP aligns a typical pair of binding sites at ∼12.5 ms using a single CPU core, ∼1 ms using 12 cores on a standard desktop machine, and performs a PDB-wide scan in 1-2 min. We perform rigorous validation of the algorithm at multiple levels of complexity and show that FLAPP provides accurate alignments. We also present a case study involving vitamin B12 binding sites to showcase the usefulness of FLAPP for performing an exhaustive alignment-based PDB-wide scan. We expect that this tool will be invaluable to the scientific community to quickly align millions of site pairs on a normal desktop machine to gain insights into protein function and drug discovery for drug target and off-target identification and polypharmacology.

GraphSite: Ligand Binding Site Classification with Deep Graph Learning

The binding of small organic molecules to protein targets is fundamental to a wide array of cellular functions. It is also routinely exploited to develop new therapeutic strategies against a variety of diseases. On that account, the ability to effectively detect and classify ligand binding sites in proteins is of paramount importance to modern structure-based drug discovery. These complex and non-trivial tasks require sophisticated algorithms from the field of artificial intelligence to achieve a high prediction accuracy. In this communication, we describe GraphSite, a deep learning-based method utilizing a graph representation of local protein structures and a state-of-the-art graph neural network to classify ligand binding sites. Using neural weighted message passing layers to effectively capture the structural, physicochemical, and evolutionary characteristics of binding pockets mitigates model overfitting and improves the classification accuracy. Indeed, comprehensive cross-validation benchmarks against a large dataset of binding pockets belonging to 14 diverse functional classes demonstrate that GraphSite yields the class-weighted F1-score of 81.7%, outperforming other approaches such as molecular docking and binding site matching. Further, it also generalizes well to unseen data with the F1-score of 70.7%, which is the expected performance in real-world applications. We also discuss new directions to improve and extend GraphSite in the future.

GrAfSS: a webserver for substructure similarity searching and comparisons in the structures of proteins and RNA

The GrAfSS (Graph theoretical Applications for Substructure Searching) webserver is a platform to search for three-dimensional substructures of: (i) amino acid side chains in protein structures; and (ii) base arrangements in RNA structures. The webserver interfaces the functions of five different graph theoretical algorithms – ASSAM, SPRITE, IMAAAGINE, NASSAM and COGNAC – into a single substructure searching suite. Users will be able to identify whether a three-dimensional (3D) arrangement of interest, such as a ligand binding site or 3D motif, observed in a protein or RNA structure can be found in other structures available in the Protein Data Bank (PDB). The webserver also allows users to determine whether a protein or RNA structure of interest contains substructural arrangements that are similar to known motifs or 3D arrangements. These capabilities allow for the functional annotation of new structures that were either experimentally determined or computationally generated (such as the coordinates generated by AlphaFold2) and can provide further insights into the diversity or conservation of functional mechanisms of structures in the PDB. The computed substructural superpositions are visualized using integrated NGL viewers. The GrAfSS server is available at http://mfrlab.org/grafss/.

Computational methods to assist in the discovery of pharmacological chaperones for rare diseases

Pharmacological chaperones are chemical compounds able to bind proteins and stabilize them against denaturation and following degradation. Some pharmacological chaperones have been approved, or are under investigation, for the treatment of rare inborn errors of metabolism, caused by genetic mutations that often can destabilize the structure of the wild-type proteins expressed by that gene. Given that, for rare diseases, there is a general lack of pharmacological treatments, many expectations are poured out on this type of compounds. However, their discovery is not straightforward. In this review, we would like to focus on the computational methods that can assist and accelerate the search for these compounds, showing also examples in which these methods were successfully applied for the discovery of promising molecules belonging to this new category of pharmacologically active compounds.

A New Strategy for Multitarget Drug Discovery/Repositioning Through the Identification of Similar 3D Amino Acid Patterns Among Proteins Structures: The Case of Tafluprost and its Effects on Cardiac Ion Channels

The identification of similar three-dimensional (3D) amino acid patterns among different proteins might be helpful to explain the polypharmacological profile of many currently used drugs. Also, it would be a reasonable first step for the design of novel multitarget compounds. Most of the current computational tools employed for this aim are limited to the comparisons among known binding sites, and do not consider several additional important 3D patterns such as allosteric sites or other conserved motifs. In the present work, we introduce Geomfinder2.0, which is a new and improved version of our previously described algorithm for the deep exploration and discovery of similar and druggable 3D patterns. As compared with the original version, substantial improvements that have been incorporated to our software allow: (i) to compare quaternary structures, (ii) to deal with a list of pairs of structures, (iii) to know how druggable is the zone where similar 3D patterns are detected and (iv) to significantly reduce the execution time. Thus, the new algorithm achieves up to 353x speedup as compared to the previous sequential version, allowing the exploration of a significant number of quaternary structures in a reasonable time. In order to illustrate the potential of the updated Geomfinder version, we show a case of use in which similar 3D patterns were detected in the cardiac ions channels NaV1.5 and TASK-1. These channels are quite different in terms of structure, sequence and function and both have been regarded as important targets for drugs aimed at treating atrial fibrillation. Finally, we describe the in vitro effects of tafluprost (a drug currently used to treat glaucoma, which was identified as a novel putative ligand of NaV1.5 and TASK-1) upon both ion channels’ activity and discuss its possible repositioning as a novel antiarrhythmic drug.

SiteMotif: A graph-based algorithm for deriving structural motifs in Protein Ligand binding sites

Studying similarities in protein molecules has become a fundamental activity in much of biology and biomedical research, for which methods such as multiple sequence alignments are widely used. Most methods available for such comparisons cater to studying proteins which have clearly recognizable evolutionary relationships but not to proteins that recognize the same or similar ligands but do not share similarities in their sequence or structural folds. In many cases, proteins in the latter class share structural similarities only in their binding sites. While several algorithms are available for comparing binding sites, there are none for deriving structural motifs of the binding sites, independent of the whole proteins. We report the development of SiteMotif, a new algorithm that compares binding sites from multiple proteins and derives sequence-order independent structural site motifs. We have tested the algorithm at multiple levels of complexity and demonstrate its performance in different scenarios. We have benchmarked against 3 current methods available for binding site comparison and demonstrate superior performance of our algorithm. We show that SiteMotif identifies new structural motifs of spatially conserved residues in proteins, even when there is no sequence or fold-level similarity. We expect SiteMotif to be useful for deriving key mechanistic insights into the mode of ligand interaction, predict the ligand type that a protein can bind and improve the sensitivity of functional annotation.

A large number of biological functions are orchestrated by proteins. The function of proteins is governed by its structure and its interacting ligand. However, it is known that not all residues are involved in ligand recognition. More specifically, residues that are located within 4.5 Å of ligand atoms are considered to be ’binding sites’. Here, we have developed an algorithm called SiteMotif that efficiently aligns multiple binding sites into a common frame. This process enables us to derive conservation among the binding site residues in a sequence order independent manner. The algorithm was validated extensively across five different levels and measured binding site similarities in each of them. Previous research has found multiple instances where different proteins have comparable binding sites and hence perform the same function. We present the ability of our method to detect such scenarios. Finally, As a use case, we applied SiteMotif to a set of glutathione binding proteins and derived a site based sequence motif characteristic of all glutathione binding proteins.

Unexpected similarity between HIV-1 reverse transcriptase and tumor necrosis factor binding sites revealed by computer vision

Rationalizing the identification of hidden similarities across the repertoire of druggable protein cavities remains a major hurdle to a true proteome-wide structure-based discovery of novel drug candidates. We recently described a new computational approach (ProCare), inspired by numerical image processing, to identify local similarities in fragment-based subpockets. During the validation of the method, we unexpectedly identified a possible similarity in the binding pockets of two unrelated targets, human tumor necrosis factor alpha (TNF-α) and HIV-1 reverse transcriptase (HIV-1 RT). Microscale thermophoresis experiments confirmed the ProCare prediction as two of the three tested and FDA-approved HIV-1 RT inhibitors indeed bind to soluble human TNF-α trimer. Interestingly, the herein disclosed similarity could be revealed neither by state-of-the-art binding sites comparison methods nor by ligand-based pairwise similarity searches, suggesting that the point cloud registration approach implemented in ProCare, is uniquely suited to identify local and unobvious similarities among totally unrelated targets.

CHARMM-GUI Ligand Designer for Template-Based Virtual Ligand Design in a Binding Site

Rational drug design involves a task of finding ligands that would bind to a specific target protein. This work presents CHARMM-GUI Ligand Designer that is an intuitive and interactive web-based tool to design virtual ligands that match the shape and chemical features of a given protein binding site. Ligand Designer provides ligand modification capabilities with 3D visualization that allow researchers to modify and redesign virtual ligands while viewing how the protein-ligand interactions are affected. Virtual ligands can also be parameterized for further molecular dynamics (MD) simulations and free energy calculations. Using 8 targets from 8 different protein classes in the directory of useful decoys, enhanced (DUD-E) data set, we show that Ligand Designer can produce similar ligands to the known active ligands in the crystal structures. Ligand Designer also produces stable protein-ligand complex structures when tested using short MD simulations. We expect that Ligand Designer can be a useful and user-friendly tool to design small molecules in any given potential ligand binding site on a protein of interest.

CHARMM-GUI LBS Finder & Refiner for Ligand Binding Site Prediction and Refinement

A protein performs its task by binding a variety of ligands in its local region that is also known as the ligand-binding-site (LBS). Therefore, accurate prediction, characterization, and refinement of LBS can facilitate protein functional annotations and structure-based drug design. In this work, we present CHARMM-GUI LBS Finder & Refiner (https://www.charmm-gui.org/input/lbsfinder) that predicts potential LBS, offers interactive features for local LBS structure analysis, and prepares various molecular dynamics (MD) systems and inputs by setting up distance restraint potentials for LBS structure refinement. LBS Finder & Refiner supports 5 different commonly used simulation programs, such as NAMD, AMBER, GROMACS, GENESIS, and OpenMM, for LBS structure refinement together with hydrogen mass repartitioning. The capability of LBS Finder & Refiner is illustrated through LBS structure predictions and refinements of 48 modeled and 20 apo benchmark target proteins. Overall, successful LBS structure predictions and refinements are seen in our benchmark tests. We hope that LBS Finder & Refiner is useful to predict, characterize, and refine potential LBS on any given protein of interest.

Ligand-Binding-Site Refinement to Generate Reliable Holo Protein Structure Conformations from Apo Structures

The first important step in a structure-based virtual screening is the judicious selection of a receptor protein. In cases where the holo protein receptor structure is unavailable, significant reduction in virtual screening performance has been reported. In this work, we present a robust method to generate reliable holo protein structure conformations from apo structures using molecular dynamics (MD) simulation with restraints derived from holo structure binding-site templates. We perform benchmark tests on two different datasets: 40 structures from a directory of useful decoy-enhanced (DUD-E) and 84 structures from the Gunasekaran dataset. Our results show successful refinement of apo binding-site structures toward holo conformations in 82% of the test cases. In addition, virtual screening performance of 40 DUD-E structures is significantly improved using our MD-refined structures as receptors with an average enrichment factor (EF), an EF_1% value of 6.2 compared to apo structures with 3.5. Docking of native ligands to the refined structures shows an average ligand root mean square deviation (RMSD) of 1.97 Å (DUD-E dataset and Gunasekaran dataset) relative to ligands in the holo crystal structures, which is comparable to the self-docking (i.e., docking of the native ligand back to its crystal structure receptor) average, 1.34 Å (DUD-E dataset) and 1.36 Å (Gunasekaran dataset). On the other hand, docking to the apo structures yields an average ligand RMSD of 3.65 Å (DUD-E) and 2.90 Å (Gunasekaran). These results indicate that our method is robust and can be useful to improve virtual screening performance of apo structures.

A systematic analysis of protein-carbohydrate interactions in the Protein Data Bank

Protein-carbohydrate interactions underlie essential biological processes. Elucidating the mechanism of protein-carbohydrate recognition is a prerequisite for modeling and optimizing protein-carbohydrate interactions, which will help in discovery of carbohydrate-derived therapeutics. In this work, we present a survey of a curated database consisting of 6,402 protein-carbohydrate complexes in the Protein Data Bank (PDB). We performed an all-against-all comparison of a subset of nonredundant binding sites, and the result indicates that the interaction pattern similarity is not completely relevant to the binding site structural similarity. Investigation of both binding site and ligand promiscuities reveals that the geometry of chemical feature points is more important than local backbone structure in determining protein-carbohydrate interactions. A further analysis on the frequency and geometry of atomic interactions shows that carbohydrate functional groups are not equally involved in binding interactions. Finally, we discuss the usefulness of protein-carbohydrate complexes in the PDB with acknowledgement that the carbohydrates in many structures are incomplete.

BionoiNet: ligand-binding site classification with off-the-shelf deep neural network

MOTIVATION

Fast and accurate classification of ligand-binding sites in proteins with respect to the class of binding molecules is invaluable not only to the automatic functional annotation of large datasets of protein structures but also to projects in protein evolution, protein engineering and drug development. Deep learning techniques, which have already been successfully applied to address challenging problems across various fields, are inherently suitable to classify ligand-binding pockets. Our goal is to demonstrate that off-the-shelf deep learning models can be employed with minimum development effort to recognize nucleotide- and heme-binding sites with a comparable accuracy to highly specialized, voxel-based methods.

RESULTS

We developed BionoiNet, a new deep learning-based framework implementing a popular ResNet model for image classification. BionoiNet first transforms the molecular structures of ligand-binding sites to 2D Voronoi diagrams, which are then used as the input to a pretrained convolutional neural network classifier. The ResNet model generalizes well to unseen data achieving the accuracy of 85.6% for nucleotide- and 91.3% for heme-binding pockets. BionoiNet also computes significance scores of pocket atoms, called BionoiScores, to provide meaningful insights into their interactions with ligand molecules. BionoiNet is a lightweight alternative to computationally expensive 3D architectures.

AVAILABILITY AND IMPLEMENTATION

BionoiNet is implemented in Python with the source code freely available at: https://github.com/CSBG-LSU/BionoiNet.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Improving detection of protein-ligand binding sites with 3D segmentation

In recent years machine learning (ML) took bio- and cheminformatics fields by storm, providing new solutions for a vast repertoire of problems related to protein sequence, structure, and interactions analysis. ML techniques, deep neural networks especially, were proven more effective than classical models for tasks like predicting binding affinity for molecular complex. In this work we investigated the earlier stage of drug discovery process - finding druggable pockets on protein surface, that can be later used to design active molecules. For this purpose we developed a 3D fully convolutional neural network capable of binding site segmentation. Our solution has high prediction accuracy and provides intuitive representations of the results, which makes it easy to incorporate into drug discovery projects. The model's source code, together with scripts for most common use-cases is freely available at http://gitlab.com/cheminfIBB/kalasanty.

Ligand-Binding-Site Structure Refinement Using Molecular Dynamics with Restraints Derived from Predicted Binding Site Templates

Accurate modeling of ligand-binding-site structures plays a critical role in structure-based virtual screening. However, the structures of the ligand-binding site in most predicted protein models are generally of low quality and need refinements. In this work, we present a ligand-binding-site structure refinement protocol using molecular dynamics simulation with restraints derived from predicted binding site templates. Our benchmark validation shows great performance for 40 diverse sets of proteins from the Astex list. The ligand-binding sites on modeled protein structures are consistently refined using our method with an average Cα RMSD improvement of 0.90 Å. Comparison of ligand binding modes from ligand docking to initial unrefined and refined structures shows an average of 1.97 Å RMSD improvement in the refined structures. These results demonstrate a promising new method of structure refinement for protein ligand-binding-site structures.

Recent applications of deep learning and machine intelligence on in silico drug discovery: methods, tools and databases

The identification of interactions between drugs/compounds and their targets is crucial for the development of new drugs. In vitro screening experiments (i.e. bioassays) are frequently used for this purpose; however, experimental approaches are insufficient to explore novel drug-target interactions, mainly because of feasibility problems, as they are labour intensive, costly and time consuming. A computational field known as 'virtual screening' (VS) has emerged in the past decades to aid experimental drug discovery studies by statistically estimating unknown bio-interactions between compounds and biological targets. These methods use the physico-chemical and structural properties of compounds and/or target proteins along with the experimentally verified bio-interaction information to generate predictive models. Lately, sophisticated machine learning techniques are applied in VS to elevate the predictive performance. The objective of this study is to examine and discuss the recent applications of machine learning techniques in VS, including deep learning, which became highly popular after giving rise to epochal developments in the fields of computer vision and natural language processing. The past 3 years have witnessed an unprecedented amount of research studies considering the application of deep learning in biomedicine, including computational drug discovery. In this review, we first describe the main instruments of VS methods, including compound and protein features (i.e. representations and descriptors), frequently used libraries and toolkits for VS, bioactivity databases and gold-standard data sets for system training and benchmarking. We subsequently review recent VS studies with a strong emphasis on deep learning applications. Finally, we discuss the present state of the field, including the current challenges and suggest future directions. We believe that this survey will provide insight to the researchers working in the field of computational drug discovery in terms of comprehending and developing novel bio-prediction methods.

Updates to Binding MOAD (Mother of All Databases): Polypharmacology Tools and Their Utility in Drug Repurposing

The goal of Binding MOAD is to provide users with a data set focused on high-quality x-ray crystal structures that have been solved with biologically relevant ligands bound. Where available, experimental binding affinities (K_a, K_d, K_i, IC₅₀) are provided from the primary literature of the crystal structure. The database has been updated regularly since 2005, and this most recent update has added nearly 7000 new structures (growth of 21%). MOAD currently contains 32,747 structures, composed of 9117 protein families and 16,044 unique ligands. The data are freely available on www.BindingMOAD.org. This paper outlines updates to the data in Binding MOAD as well as improvements made to both the website and its contents. The NGL viewer has been added to improve visualization of the ligands and protein structures. MarvinJS has been implemented, over the outdated MarvinView, to work with JChem for small molecule searching in the database. To add tools for predicting polypharmacology, we have added information about sequence, binding-site, and ligand similarity between entries in the database. A main premise behind polypharmacology is that similar binding sites will bind similar ligands. The large amount of protein-ligand information available in Binding MOAD allows us to compute pairwise ligand and binding-site similarities. Lists of similar ligands and similar binding sites have been added to allow users to identify potential polypharmacology pairs. To show the utility of the polypharmacology data, we detail a few examples from Binding MOAD of drug repurposing targets with their respective similarities.

Stalis: A Computational Method for Template-Based Ab Initio Ligand Design

Proteins interact with small molecules through specific molecular recognition, which is central to essential biological functions in living systems. Therefore, understanding such interactions is crucial for basic sciences and drug discovery. Here, we present Structure template-based ab initio ligand design solution (Stalis), a knowledge-based approach that uses structure templates from the Protein Data Bank libraries of whole ligands and their fragments and generates a set of molecules (virtual ligands) whose structures represent the pocket shape and chemical features of a given target binding site. Our benchmark performance evaluation shows that ligand structure-based virtual screening using virtual ligands from Stalis outperforms a receptor structure-based virtual screening using AutoDock Vina, demonstrating reliable overall screening performance applicable to computational high-throughput screening. However, virtual ligands from Stalis are worse in recognizing active compounds at the small fraction of a rank-ordered list of screened library compounds than crystal ligands, due to the low resolution of the virtual ligand structures. In conclusion, Stalis can facilitate drug discovery research by designing virtual ligands that can be used for fast ligand structure-based virtual screening. Moreover, Stalis provides actual three-dimensional ligand structures that likely bind to a target protein, enabling to gain structural insight into potential ligands. Stalis can be an efficient computational platform for high-throughput ligand design for fundamental biological study and drug discovery research at the proteomic level. © 2019 Wiley Periodicals, Inc.

DeepDrug3D: Classification of ligand-binding pockets in proteins with a convolutional neural network

Comprehensive characterization of ligand-binding sites is invaluable to infer molecular functions of hypothetical proteins, trace evolutionary relationships between proteins, engineer enzymes to achieve a desired substrate specificity, and develop drugs with improved selectivity profiles. These research efforts pose significant challenges owing to the fact that similar pockets are commonly observed across different folds, leading to the high degree of promiscuity of ligand-protein interactions at the system-level. On that account, novel algorithms to accurately classify binding sites are needed. Deep learning is attracting a significant attention due to its successful applications in a wide range of disciplines. In this communication, we present DeepDrug3D, a new approach to characterize and classify binding pockets in proteins with deep learning. It employs a state-of-the-art convolutional neural network in which biomolecular structures are represented as voxels assigned interaction energy-based attributes. The current implementation of DeepDrug3D, trained to detect and classify nucleotide- and heme-binding sites, not only achieves a high accuracy of 95%, but also has the ability to generalize to unseen data as demonstrated for steroid-binding proteins and peptidase enzymes. Interestingly, the analysis of strongly discriminative regions of binding pockets reveals that this high classification accuracy arises from learning the patterns of specific molecular interactions, such as hydrogen bonds, aromatic and hydrophobic contacts. DeepDrug3D is available as an open-source program at https://github.com/pulimeng/DeepDrug3D with the accompanying TOUGH-C1 benchmarking dataset accessible from https://osf.io/enz69/.

Prediction Methods of Herbal Compounds in Chinese Medicinal Herbs

Chinese herbal medicine has recently gained worldwide attention. The curative mechanism of Chinese herbal medicine is compared with that of western medicine at the molecular level. The treatment mechanism of most Chinese herbal medicines is still not clear. How do we integrate Chinese herbal medicine compounds with modern medicine? Chinese herbal medicine drug-like prediction method is particularly important. A growing number of Chinese herbal source compounds are now widely used as drug-like compound candidates. An important way for pharmaceutical companies to develop drugs is to discover potentially active compounds from related herbs in Chinese herbs. The methods for predicting the drug-like properties of Chinese herbal compounds include the virtual screening method, pharmacophore model method and machine learning method. In this paper, we focus on the prediction methods for the medicinal properties of Chinese herbal medicines. We analyze the advantages and disadvantages of the above three methods, and then introduce the specific steps of the virtual screening method. Finally, we present the prospect of the joint application of various methods.

G-LoSA for Prediction of Protein-Ligand Binding Sites and Structures

Recent advances in high-throughput structure determination and computational protein structure prediction have significantly enriched the universe of protein structure. However, there is still a large gap between the number of available protein structures and that of proteins with annotated function in high accuracy. Computational structure-based protein function prediction has emerged to reduce this knowledge gap. The identification of a ligand binding site and its structure is critical to the determination of a protein's molecular function. We present a computational methodology for predicting small molecule ligand binding site and ligand structure using G-LoSA, our protein local structure alignment and similarity measurement tool. All the computational procedures described here can be easily implemented using G-LoSA Toolkit, a package of standalone software programs and preprocessed PDB structure libraries. G-LoSA and G-LoSA Toolkit are freely available to academic users at http://compbio.lehigh.edu/GLoSA . We also illustrate a case study to show the potential of our template-based approach harnessing G-LoSA for protein function prediction.