Improving protein-ligand binding site prediction accuracy by classification of inner pocket points using local features

Preclinical alternative drug discovery programs for monogenic rare diseases. Should small molecules or gene therapy be used? The case of hereditary spastic paraplegias

Patients diagnosed with rare diseases and their and families search desperately to organize drug discovery campaigns. Alternative models that differ from default paradigms offer real opportunities. There are, however, no clear guidelines for the development of such models, which reduces success rates and raises costs. We address the main challenges in making the discovery of new preclinical treatments more accessible, using rare hereditary paraplegia as a paradigmatic case. First, we discuss the necessary expertise, and the patients' clinical and genetic data. Then, we revisit gene therapy, de novo drug development, and drug repurposing, discussing their applicability. Moreover, we explore a pool of recommended in silico tools for pathogenic variant and protein structure prediction, virtual screening, and experimental validation methods, discussing their strengths and weaknesses. Finally, we focus on successful case applications.

Ribosomal translation of fluorinated non-canonical amino acids for de novo biologically active fluorinated macrocyclic peptides

Fluorination has emerged as a promising strategy in medicinal chemistry to improve the pharmacological profiles of drug candidates. Similarly, incorporating fluorinated non-canonical amino acids into macrocyclic peptides expands chemical diversity and enhances their pharmacological properties, from improved metabolic stability to enhanced cell permeability and target interactions. However, only a limited number of fluorinated non-canonical amino acids, which are canonical amino acid analogs, have been incorporated into macrocyclic peptides by ribosomes for de novo construction and target-based screening of fluorinated macrocyclic peptides. In this study, we report the ribosomal translation of a series of distinct fluorinated non-canonical amino acids, including mono-to tri-fluorinated variants, as well as fluorinated l-amino acids, d-amino acids, β-amino acids, etc. This enabled the de novo discovery of fluorinated macrocyclic peptides with high affinity for EphA2, and particularly the identification of those exhibiting broad-spectrum activity against Gram-negative bacteria by targeting the BAM complex. This study not only expands the scope of ribosomally translatable fluorinated amino acids but also underscores the versatility of fluorinated macrocyclic peptides as potent therapeutic agents.

DEAttentionDTA: protein-ligand binding affinity prediction based on dynamic embedding and self-attention

MOTIVATION

Predicting protein-ligand binding affinity is crucial in new drug discovery and development. However, most existing models rely on acquiring 3D structures of elusive proteins. Combining amino acid sequences with ligand sequences and better highlighting active sites are also significant challenges.

RESULTS

We propose an innovative neural network model called DEAttentionDTA, based on dynamic word embeddings and a self-attention mechanism, for predicting protein-ligand binding affinity. DEAttentionDTA takes the 1D sequence information of proteins as input, including the global sequence features of amino acids, local features of the active pocket site, and linear representation information of the ligand molecule in the SMILE format. These three linear sequences are fed into a dynamic word-embedding layer based on a 1D convolutional neural network for embedding encoding and are correlated through a self-attention mechanism. The output affinity prediction values are generated using a linear layer. We compared DEAttentionDTA with various mainstream tools and achieved significantly superior results on the same dataset. We then assessed the performance of this model in the p38 protein family.

AVAILABILITY AND IMPLEMENTATION

The resource codes are available at https://github.com/whatamazing1/DEAttentionDTA.

Structure-Based Discovery of Allosteric Inhibitors Targeting a New Druggable Site in the Respiratory Syncytial Virus Polymerase

Respiratory syncytial virus (RSV) is a major cause of severe lower respiratory infections for which effective treatment options remain limited. Herein, we employed a computational structure-based design strategy aimed at identifying potential targets for a new class of allosteric inhibitors. Our investigation led to the discovery of a previously undisclosed allosteric binding site within the RSV polymerase, the large (L) protein. This discovery was achieved through a combination of virtual screening and molecular dynamics simulations. Subsequently, we identified two inhibitors, 6a and 10b, which both exhibited promising antiviral activity in the low micromolar range. Resistance profiling revealed a distinctive pattern in how RSV evaded treatment with this class of inhibitors. This pattern strongly suggested that this class of small molecules was targeting a new binding site in the RSV L protein, aligning with the computational predictions made in our study. This study paves the way for the development of more potent inhibitors for combating RSV infections by targeting a new druggable pocket within the RdRp which does not overlap with previously known resistance sites.

ANN based prediction of ligand binding sites outside deep cavities to facilitate drug designing

The ever-changing environmental conditions and pollution are the prime reasons for the onset of several emerging and re-merging diseases. This demands the faster designing of new drugs to curb the deadly diseases in less waiting time to cure the animals and humans. Drug molecules interact with only protein surface on specific locations termed as ligand binding sites (LBS). Therefore, the knowledge of LBS is required for rational drug designing. Existing geometrical LBS prediction methods rely on search of cavities based on the fact that 83% of the LBS found in deep cavities, however, these methods usually fail where LBS localize outside deep cavities. To overcome this challenge, the present work provides an artificial neural network (ANN) based method to predict LBS outside deep cavities in animal proteins including human to facilitate drug designing. In the present work a feed-forward backpropagation neural network was trained by utilizing 38 structural, atomic, physiochemical, and evolutionary discriminant features of LBS and non-LBS residues localized in the extracted roughest patch on protein surface. The performance of this ANN based prediction method was found 76% better for those proteins where cavity subspace (extracted by MetaPocket 2.0, a consensus method) failed to predict LBS due to their localization outside the deep cavities. The prediction of LBS outside deep cavities will facilitate in drug designing for the proteins where it is not possible due to lack of LBS information as the geometrical LBS prediction methods rely on extraction of deep cavities.

Graph Attention Site Prediction (GrASP): Identifying Druggable Binding Sites Using Graph Neural Networks with Attention

Identifying and discovering druggable protein binding sites is an important early step in computer-aided drug discovery, but it remains a difficult task where most campaigns rely on a priori knowledge of binding sites from experiments. Here, we present a binding site prediction method called Graph Attention Site Prediction (GrASP) and re-evaluate assumptions in nearly every step in the site prediction workflow from data set preparation to model evaluation. GrASP is able to achieve state-of-the-art performance at recovering binding sites in PDB structures while maintaining a high degree of precision which will minimize wasted computation in downstream tasks such as docking and free energy perturbation.

Searching for new plastic-degrading enzymes from the plastisphere of alpine soils using a metagenomic mining approach

Plastic materials, including microplastics, accumulate in all types of ecosystems, even in remote and cold environments such as the European Alps. This pollution poses a risk for the environment and humans and needs to be addressed. Using shotgun DNA metagenomics of soils collected in the eastern Swiss Alps at about 3,000 m a.s.l., we identified genes and their proteins that potentially can degrade plastics. We screened the metagenomes of the plastisphere and the bulk soil with a differential abundance analysis, conducted similarity-based screening with specific databases dedicated to putative plastic-degrading genes, and selected those genes with a high probability of signal peptides for extracellular export and a high confidence for functional domains. This procedure resulted in a final list of nine candidate genes. The lengths of the predicted proteins were between 425 and 845 amino acids, and the predicted genera producing these proteins belonged mainly to Caballeronia and Bradyrhizobium. We applied functional validation, using heterologous expression followed by enzymatic assays of the supernatant. Five of the nine proteins tested showed significantly increased activities when we used an esterase assay, and one of these five proteins from candidate genes, a hydrolase-type esterase, clearly had the highest activity, by more than double. We performed the fluorescence assays for plastic degradation of the plastic types BI-OPL and ecovio® only with proteins from the five candidate genes that were positively active in the esterase assay, but like the negative controls, these did not show any significantly increased activity. In contrast, the activity of the positive control, which contained a PLA-degrading gene insert known from the literature, was more than 20 times higher than that of the negative controls. These findings suggest that in silico screening followed by functional validation is suitable for finding new plastic-degrading enzymes. Although we only found one new esterase enzyme, our approach has the potential to be applied to any type of soil and to plastics in various ecosystems to search rapidly and efficiently for new plastic-degrading enzymes.

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.

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.

Open-Source Machine Learning in Computational Chemistry

The field of computational chemistry has seen a significant increase in the integration of machine learning concepts and algorithms. In this Perspective, we surveyed 179 open-source software projects, with corresponding peer-reviewed papers published within the last 5 years, to better understand the topics within the field being investigated by machine learning approaches. For each project, we provide a short description, the link to the code, the accompanying license type, and whether the training data and resulting models are made publicly available. Based on those deposited in GitHub repositories, the most popular employed Python libraries are identified. We hope that this survey will serve as a resource to learn about machine learning or specific architectures thereof by identifying accessible codes with accompanying papers on a topic basis. To this end, we also include computational chemistry open-source software for generating training data and fundamental Python libraries for machine learning. Based on our observations and considering the three pillars of collaborative machine learning work, open data, open source (code), and open models, we provide some suggestions to the community.

xProtCAS: A Toolkit for Extracting Conserved Accessible Surfaces from Protein Structures

The identification of protein surfaces required for interaction with other biomolecules broadens our understanding of protein function, their regulation by post-translational modification, and the deleterious effect of disease mutations. Protein interaction interfaces are often identifiable as patches of conserved residues on a protein's surface. However, finding conserved accessible surfaces on folded regions requires an understanding of the protein structure to discriminate between functional and structural constraints on residue conservation. With the emergence of deep learning methods for protein structure prediction, high-quality structural models are now available for any protein. In this study, we introduce tools to identify conserved surfaces on AlphaFold2 structural models. We define autonomous structural modules from the structural models and convert these modules to a graph encoding residue topology, accessibility, and conservation. Conserved surfaces are then extracted using a novel eigenvector centrality-based approach. We apply the tool to the human proteome identifying hundreds of uncharacterised yet highly conserved surfaces, many of which contain clinically significant mutations. The xProtCAS tool is available as open-source Python software and an interactive web server.

Pharmacoinformatic Investigation of Silymarin as a Potential Inhibitor against Nemopilema nomurai Jellyfish Metalloproteinase Toxin-like Protein

Jellyfish stings pose a major threat to swimmers and fishermen worldwide. These creatures have explosive cells containing one large secretory organelle called a nematocyst in their tentacles, which contains venom used to immobilize prey. Nemopilema nomurai, a venomous jellyfish belonging to the phylum Cnidaria, produces venom (NnV) comprising various toxins known for their lethal effects on many organisms. Of these toxins, metalloproteinases (which belong to the toxic protease family) play a significant role in local symptoms such as dermatitis and anaphylaxis, as well as systemic reactions such as blood coagulation, disseminated intravascular coagulation, tissue injury, and hemorrhage. Hence, a potential metalloproteinase inhibitor (MPI) could be a promising candidate for reducing the effects of venom toxicity. For this study, we retrieved the Nemopilema nomurai venom metalloproteinase sequence (NnV-MPs) from transcriptome data and modeled its three-dimensional structure using AlphaFold2 in a Google Colab notebook. We employed a pharmacoinformatics approach to screen 39 flavonoids and identify the most potent inhibitor against NnV-MP. Previous studies have demonstrated the efficacy of flavonoids against other animal venoms. Based on our analysis, Silymarin emerged as the top inhibitor through ADMET, docking, and molecular dynamics analyses. In silico simulations provide detailed information on the toxin and ligand binding affinity. Our results demonstrate that Silymarin's strong inhibitory effect on NnV-MP is driven by hydrophobic affinity and optimal hydrogen bonding. These findings suggest that Silymarin could serve as an effective inhibitor of NnV-MP, potentially reducing the toxicity associated with jellyfish envenomation.

Binding Site Detection Remastered: Enabling Fast, Robust, and Reliable Binding Site Detection and Descriptor Calculation with DoGSite3

Binding site prediction on protein structures is a crucial step in early phase drug discovery whenever experimental or predicted structure models are involved. DoGSite belongs to the widely used tools for this task. It is a grid-based method that uses a Difference-of-Gaussian filter to detect cavities on the protein surface. We recently reimplemented the first version of this method, released in 2010, focusing on improved binding site detection in the presence of ligands and optimized parameters for more robust, reliable, and fast predictions and binding site descriptor calculations. Here, we introduce the new version, DoGSite3, compare it to its predecessor, and re-evaluate DoGSite on published data sets for a large-scale comparative performance evaluation.

Docking and Molecular Dynamics Identify Leads against 5 Alpha Reductase 2 for Benign Prostate Hyperplasia Treatment

Steroid 5 alpha-reductase 2 (5αR-2) is a membrane-embedded protein that together with other isoforms plays a key role in the metabolism of steroids. This enzyme catalyzes the reduction of testosterone to the more potent ligand, dihydrotestosterone (DHT) in the prostate. Androgens, testosterone, and DHT play important roles in prostate growth, development, and function. At the same time, both testosterone and DHT have been implicated in the pathogenesis of benign prostate hyperplasia (BPH). Inhibition of the DHT formation, therefore, provides a therapeutic strategy that offers the possibility of preventing, delaying, or treating BPH. Currently, two steroidal drugs that inhibit 5αR-2, dutasteride and finasteride, have been approved for clinical use. These two come at a high cost and also portray undesirable sexual side effects which necessitate the need to find new chemotherapeutic alternatives for the disease. Based on the aforementioned, finasteride and dutasteride were subjected to scaffold hopping, fragment-based de novo design, molecular docking, and molecular dynamics simulations employing databases like ChEMBL, DrugBank, PubChem, ChemSpider, and Zinc15 in the identification of potential hits targeting 5αR-2. Altogether, ten novel compounds targeting 5αR-2 were identified with binding energies lower or comparable to finasteride and dutasteride, the main inhibitors for this target. Molecular docking and molecular dynamics simulations studies identify amino acid residues Glu57, Phe219, Phe223, and Leu224 to be critical for ligand binding and complex stability. The physicochemical and pharmacological profiling suggests the potential of the hit compounds to be drug-like and orally active. Similarly, the quality parameter assessments revealed the hits possess LELP greater than 3 implying their promise as lead-like molecules. The compounds A5, A9, and A10 were, respectively, predicted to treat prostate disorders with Pa (0.188, 0.361, and 0.270) and Pi (0.176, 0.050, and 0.093), while A8 and A9 were found to be associated with BPH treatment with Pa (0.09 and 0.127) and Pi (0.077 and 0.033), respectively. Structural similarity searches via DrugBank identified the drugs faropenem, acemetacin, estradiol valerate, and yohimbine to be useful for BPH treatment suggesting the de novo designed ligands as potential chemotherapeutic agents for treating this disease.

Synthesis, modification and application of fish skin gelatin-based hydrogel as sustainable and versatile bioresource of antidiabetic peptide

Gelatin hydrogel is widely employed in various fields, however, commercially available gelatin hydrogels are mostly derived from mammalian which has many disadvantages due to the supply and ethical issues. In this study, the properties of hydrogels from fish-derived collagen fabricated with varying Glutaraldehyde (GA) determined. The antidiabetic properties of salmon gelatin (SG) and tilapia gelatin (TG) was also evaluated against α-glucosidase. Glutaraldehyde-crosslinked salmon gelatin and tilapia gelatin were used, and compared with different concentrations of GA by 0.05 %, 0.1 %, and 0.15 %. Water absorbency, swelling, porosity, pore size and water retention of the hydrogels were dependent on the degree of crosslinking. The synthesis of hydrogels was confirmed by FTIR study. Scanning electron microscope (SEM) observation showed that all hydrogels have a porous structure with irregular shapes and heterogeneous morphology. Performance tests showed that gelatin-GA 0.05 % mixture had the best performance. Antidiabetic bioactivity in vitro and in silico tests showed that the active peptides of SG and TG showed a high binding affinity to α-glucosidase enzyme. In conclusion, SG and TG cross-linked GA 0.05 % have the potential as an antidiabetic agent and as a useful option over mammalian-derived gelatin.

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.

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.

Probabilistic Pocket Druggability Prediction via One-Class Learning

The choice of target pocket is a key step in a drug discovery campaign. This step can be supported by in silico druggability prediction. In the literature, druggability prediction is often approached as a two-class classification task that distinguishes between druggable and non-druggable (or less druggable) pockets (or voxels). Apart from obvious cases, however, the non-druggable class is conceptually ambiguous. This is because any pocket (or target) is only non-druggable until a drug is found for it. It is therefore more appropriate to adopt a one-class approach, which uses only unambiguous information, namely, druggable pockets. Here, we propose using the import vector domain description (IVDD) algorithm to support this task. IVDD is a one-class probabilistic kernel machine that we previously introduced. To feed the algorithm, we use customized DrugPred descriptors computed via NanoShaper. Our results demonstrate the feasibility and effectiveness of the approach. In particular, we can remove or mitigate biases chiefly due to the labeling.

Consecutive Aromatic Residues Are Required for Improved Efficacy of β-Sheet Breakers

Alzheimer’s disease is a fatal neurodegenerative malady which up to very recently did not have approved therapy modifying its course. After controversial approval of aducanumab (monoclonal antibody clearing β-amyloid plaques) by FDA for use in very early stages of disease, possibly new avenue opened for the treatment of patients. In line with this approach is search for compounds blocking aggregation into amyloid oligomers subsequently forming fibrils or compounds helping in getting rid of plaques formed by β-amyloid fibrils. Here we present in silico work on 627 sixtapeptide β-sheet breakers (BSBs) containing consecutive three aromatic residues. Three of these BSBs caused dissociation of one or two β-amyloid chains from U-shaped β-amyloid protofibril model 2BEG after docking and subsequent molecular dynamics simulations. Thorough analysis of our results let us postulate that the first steps of binding these successful BSBs involve π–π interactions with stacked chains of F19 and later also with F20 (F3 and F4 in 2BEG model of protofibril). The consecutive location of aromatic residues in BSBs makes them more attractive for chains of stacked F3 and F4 within the 2BEG model. Spotted by us, BSBs may be prospective lead compounds for an anti-Alzheimer’s therapy.

BiRDS - Binding Residue Detection from Protein Sequences Using Deep ResNets

Protein-drug interactions play important roles in many biological processes and therapeutics. Predicting the binding sites of a protein helps to discover such interactions. New drugs can be designed to optimize these interactions, improving protein function. The tertiary structure of a protein decides the binding sites available to the drug molecule, but the determination of the 3D structure is slow and expensive. Conversely, the determination of the amino acid sequence is swift and economical. Although quick and accurate prediction of the binding site using just the sequence is challenging, the application of Deep Learning, which has been hugely successful in several biochemical tasks, makes it feasible. BiRDS is a Residual Neural Network that predicts the protein's most active binding site using sequence information. SC-PDB, an annotated database of druggable binding sites, is used for training the network. Multiple Sequence Alignments of the proteins in the database are generated using DeepMSA, and features such as Position-Specific Scoring Matrix, Secondary Structure, and Relative Solvent Accessibility are extracted. During training, a weighted binary cross-entropy loss function is used to counter the substantial imbalance in the two classes of binding and nonbinding residues. A novel test set SC6K is introduced to compare binding-site prediction methods. BiRDS achieves an AUROC score of 0.87, and the center of 25% of its predicted binding sites lie within 4 Å of the center of the actual binding site.

Structural determinants of peptide-dependent TAP1-TAP2 transit passage targeted by viral proteins and altered by cancer-associated mutations

The TAP1-TAP2 complex transports antigenic peptide substrates into the endoplasmic reticulum (ER). In ER, the peptides are further processed and loaded on the major histocompatibility class (MHC) I molecules by the peptide loading complex (PLC). The TAP transporters are linked with the PLC; a target for cancers and viral immune evasion. But the mechanisms whereby the cancer-derived mutations in TAP1-TAP2 or viral factors targeting the PLC, interfere peptide transport are only emerging. This study describes that transit of peptides through TAP can take place via two different channels (4 or 8 helices) depending on peptide length and sequence. Molecular dynamics and binding affinity predictions of peptide-transporters demonstrated that smaller peptides (8-10 mers; e.g. AAGIGILTV, SIINFEKL) can transport quickly through the transport tunnel compared to longer peptides (15-mer; e.g. ENPVVHFFKNIVTPR). In line with a regulated and selective peptide transport by TAPs, the immunopeptidome upon IFN-γ treatment in melanoma cells induced the shorter length (9-mer) peptide presentation over MHC-I that exhibit a relatively weak binding affinity with TAP. A conserved distance between N and C terminus residues of the studied peptides in the transport tunnel were reported. Furthermore, by adversely interacting with the TAP transport passage or affecting TAPNBD domains tilt movement, the viral proteins and cancer-derived mutations in TAP1-TAP2 may induce allosteric effects in TAP that block conformation of the tunnel (closed towards ER lumen). Interestingly, some cancer-associated mutations (e.g. TAP1R372Q and TAP2R373H) can specifically interfere with selective transport channels (i.e. for longer-peptides). These results provide a model for how viruses and cancer-associated mutations targeting TAP interfaces can affect MHC-I antigen presentation, and how the IFN-γ pathway alters MHC-I antigen presentation via the kinetics of peptide transport.

DeepPocket: Ligand Binding Site Detection and Segmentation using 3D Convolutional Neural Networks

A structure-based drug design pipeline involves the development of potential drug molecules or ligands that form stable complexes with a given receptor at its binding site. A prerequisite to this is finding druggable and functionally relevant binding sites on the 3D structure of the protein. Although several methods for detecting binding sites have been developed beforehand, a majority of them surprisingly fail in the identification and ranking of binding sites accurately. The rapid adoption and success of deep learning algorithms in various sections of structural biology beckons the usage of such algorithms for accurate binding site detection. As a combination of geometry based software and deep learning, we report a novel framework, DeepPocket that utilizes 3D convolutional neural networks for the rescoring of pockets identified by Fpocket and further segments these identified cavities on the protein surface. Apart from this, we also propose another data set SC6K containing protein structures submitted in the Protein Data Bank (PDB) from January 1st, 2018, until February 28th, 2020, for ligand binding site (LBS) detection. DeepPocket's results on various binding site data sets and SC6K highlight its better performance over current state-of-the-art methods and good generalization ability over novel structures.

Finding Druggable Sites in Proteins Using TACTICS

Structure-based drug discovery efforts require knowledge of where drug-binding sites are located on target proteins. To address the challenge of finding druggable sites, we developed a machine-learning algorithm called TACTICS (trajectory-based analysis of conformations to identify cryptic sites), which uses an ensemble of molecular structures (such as molecular dynamics simulation data) as input. First, TACTICS uses k-means clustering to select a small number of conformations that represent the overall conformational heterogeneity of the data. Then, TACTICS uses a random forest model to identify potentially bindable residues in each selected conformation, based on protein motion and geometry. Lastly, residues in possible binding pockets are scored using fragment docking. As proof-of-principle, TACTICS was applied to the analysis of simulations of the SARS-CoV-2 main protease and methyltransferase and the Yersinia pestis aryl carrier protein. Our approach recapitulates known small-molecule binding sites and predicts the locations of sites not previously observed in experimentally determined structures. The TACTICS code is available at https://github.com/Albert-Lau-Lab/tactics_protein_analysis.

Synthesis, Molecular Docking Screening and Anti-Proliferative Potency Evaluation of Some New Imidazo[2,1-b]Thiazole Linked Thiadiazole Conjugates

BACKGROUND

Imidazo[2,1-b]thiazole scaffolds were reported to possess various pharmaceutical activities.

RESULTS

The novel compound named methyl-2-(1-(3-methyl-6-(p-tolyl)imidazo[2,1-b]thiazol-2-yl)ethylidene)hydrazine-1-carbodithioate 3 acted as a predecessor molecule for the synthesis of new thiadiazole derivatives incorporating imidazo[2,1-b]thiazole moiety. The reaction of 3 with the appropriate hydrazonoyl halide derivatives 4a-j and 7-9 had produced the respective 1,3,4-thiadiazole derivatives 6a-j and 10-12. The chemical composition of all the newly synthesized derivatives were confirmed by their microanalytical and spectral data (FT-IR, mass spectrometry, ¹H-NMR and ¹³C-NMR). All the produced novel compounds were screened for their anti-proliferative efficacy on hepatic cancer cell lines (HepG₂). In addition, a computational molecular docking study was carried out to determine the ability of the synthesized thiadiazole molecules to interact with active site of the target Glypican-3 protein (GPC-3). Moreover, the physiochemical properties of the synthesized compounds were derived to determine the viability of the compounds as drug candidates for hepatic cancer.

CONCLUSION

All the tested compounds had exhibited good anti-proliferative efficacy against hepatic cancer cell lines. In addition, the molecular docking results showed strong binding interactions of the synthesized compounds with the target GPC-3 protein with lower energy scores. Thus, such novel compounds may act as promising candidates as drugs against hepatocellular carcinoma.

Exploring the computational methods for protein-ligand binding site prediction

Proteins participate in various essential processes in vivo via interactions with other molecules. Identifying the residues participating in these interactions not only provides biological insights for protein function studies but also has great significance for drug discoveries. Therefore, predicting protein-ligand binding sites has long been under intense research in the fields of bioinformatics and computer aided drug discovery. In this review, we first introduce the research background of predicting protein-ligand binding sites and then classify the methods into four categories, namely, 3D structure-based, template similarity-based, traditional machine learning-based and deep learning-based methods. We describe representative algorithms in each category and elaborate on machine learning and deep learning-based prediction methods in more detail. Finally, we discuss the trends and challenges of the current research such as molecular dynamics simulation based cryptic binding sites prediction, and highlight prospective directions for the near future.

New Binding Sites, New Opportunities for GPCR Drug Discovery

Many central biological events rely on protein-ligand interactions. The identification and characterization of protein-binding sites for ligands are crucial for the understanding of functions of both endogenous ligands and synthetic drug molecules. G protein-coupled receptors (GPCRs) typically detect extracellular signal molecules on the cell surface and transfer these chemical signals across the membrane, inducing downstream cellular responses via G proteins or β-arrestin. GPCRs mediate many central physiological processes, making them important targets for modern drug discovery. Here, we focus on the most recent breakthroughs in finding new binding sites and binding modes of GPCRs and their potentials for the development of new medicines.

P2Rank: machine learning based tool for rapid and accurate prediction of ligand binding sites from protein structure

Background

Ligand binding site prediction from protein structure has many applications related to elucidation of protein function and structure based drug discovery. It often represents only one step of many in complex computational drug design efforts. Although many methods have been published to date, only few of them are suitable for use in automated pipelines or for processing large datasets. These use cases require stability and speed, which disqualifies many of the recently introduced tools that are either template based or available only as web servers.

Results

We present P2Rank, a stand-alone template-free tool for prediction of ligand binding sites based on machine learning. It is based on prediction of ligandability of local chemical neighbourhoods that are centered on points placed on the solvent accessible surface of a protein. We show that P2Rank outperforms several existing tools, which include two widely used stand-alone tools (Fpocket, SiteHound), a comprehensive consensus based tool (MetaPocket 2.0), and a recent deep learning based method (DeepSite). P2Rank belongs to the fastest available tools (requires under 1 s for prediction on one protein), with additional advantage of multi-threaded implementation.

Conclusions

P2Rank is a new open source software package for ligand binding site prediction from protein structure. It is available as a user-friendly stand-alone command line program and a Java library. P2Rank has a lightweight installation and does not depend on other bioinformatics tools or large structural or sequence databases. Thanks to its speed and ability to make fully automated predictions, it is particularly well suited for processing large datasets or as a component of scalable structural bioinformatics pipelines.

Electronic supplementary material

The online version of this article (10.1186/s13321-018-0285-8) contains supplementary material, which is available to authorized users.

Polypharmacology Within the Full Kinome: a Machine Learning Approach

Protein kinases generate nearly a thousand different protein products and regulate the majority of cellular pathways and signal transduction. It is therefore not surprising that the deregulation of kinases has been implicated in many disease states. In fact, kinase inhibitors are the largest class of new cancer therapies. Understanding polypharmacology within the full kinome, how drugs interact with many different kinases, would allow for the development of safer and more efficacious cancer therapies. A full understanding of these interactions is not experimentally feasible making highly accurate computational predictions extremely useful and important. This work aims at making a machine learning model useful for investigating the full kinome. We evaluate many feature sets for our model and get better performance over molecular docking with all of them. We demonstrate that you can achieve a nearly 60% increase in success rate at identifying binding compounds using our model over molecular docking scores.

Investigating the Importance of the Pocket-estimation Method in Pocket-based Approaches: An Illustration Using Pocket-ligand Classification

Small molecules interact with their protein target on surface cavities known as binding pockets. Pocket-based approaches are very useful in all of the phases of drug design. Their first step is estimating the binding pocket based on protein structure. The available pocket-estimation methods produce different pockets for the same target. The aim of this work is to investigate the effects of different pocket-estimation methods on the results of pocket-based approaches. We focused on the effect of three pocket-estimation methods on a pocket-ligand (PL) classification. This pocket-based approach is useful for understanding the correspondence between the pocket and ligand spaces and to develop pharmacological profiling models. We found pocket-estimation methods yield different binding pockets in terms of boundaries and properties. These differences are responsible for the variation in the PL classification results that can have an impact on the detected correspondence between pocket and ligand profiles. Thus, we highlighted the importance of the pocket-estimation method choice in pocket-based approaches.

Can We Rely on Computational Predictions To Correctly Identify Ligand Binding Sites on Novel Protein Drug Targets? Assessment of Binding Site Prediction Methods and a Protocol for Validation of Predicted Binding Sites

In the field of medicinal chemistry there is increasing focus on identifying key proteins whose biochemical functions can firmly be linked to serious diseases. Such proteins become targets for drug or inhibitor molecules that could treat or halt the disease through therapeutic action or by blocking the protein function respectively. The protein must be targeted at the relevant biologically active site for drug or inhibitor binding to be effective. As insufficient experimental data is available to confirm the biologically active binding site for novel protein targets, researchers often rely on computational prediction methods to identify binding sites. Presented herein is a short review on structure-based computational methods that (i) predict putative binding sites and (ii) assess the druggability of predicted binding sites on protein targets. This review briefly covers the principles upon which these methods are based, where they can be accessed and their reliability in identifying the correct binding site on a protein target. Based on this review, we believe that these methods are useful in predicting putative binding sites, but as they do not account for the dynamic nature of protein-ligand binding interactions, they cannot definitively identify the correct site from a ranked list of putative sites. To overcome this shortcoming, we strongly recommend using molecular docking to predict the most likely protein-ligand binding site(s) and mode(s), followed by molecular dynamics simulations and binding thermodynamics calculations to validate the docking results. This protocol provides a valuable platform for experimental and computational efforts to design novel drugs and inhibitors that target disease-related proteins.

Use of machine learning approaches for novel drug discovery

INTRODUCTION

The use of computational tools in the early stages of drug development has increased in recent decades. Machine learning (ML) approaches have been of special interest, since they can be applied in several steps of the drug discovery methodology, such as prediction of target structure, prediction of biological activity of new ligands through model construction, discovery or optimization of hits, and construction of models that predict the pharmacokinetic and toxicological (ADMET) profile of compounds.

AREAS COVERED

This article presents an overview on some applications of ML techniques in drug design. These techniques can be employed in ligand-based drug design (LBDD) and structure-based drug design (SBDD) studies, such as similarity searches, construction of classification and/or prediction models of biological activity, prediction of secondary structures and binding sites docking and virtual screening.

EXPERT OPINION

Successful cases have been reported in the literature, demonstrating the efficiency of ML techniques combined with traditional approaches to study medicinal chemistry problems. Some ML techniques used in drug design are: support vector machine, random forest, decision trees and artificial neural networks. Currently, an important application of ML techniques is related to the calculation of scoring functions used in docking and virtual screening assays from a consensus, combining traditional and ML techniques in order to improve the prediction of binding sites and docking solutions.

Proteins and Their Interacting Partners: An Introduction to Protein-Ligand Binding Site Prediction Methods

Elucidating the biological and biochemical roles of proteins, and subsequently determining their interacting partners, can be difficult and time consuming using in vitro and/or in vivo methods, and consequently the majority of newly sequenced proteins will have unknown structures and functions. However, in silico methods for predicting protein-ligand binding sites and protein biochemical functions offer an alternative practical solution. The characterisation of protein-ligand binding sites is essential for investigating new functional roles, which can impact the major biological research spheres of health, food, and energy security. In this review we discuss the role in silico methods play in 3D modelling of protein-ligand binding sites, along with their role in predicting biochemical functionality. In addition, we describe in detail some of the key alternative in silico prediction approaches that are available, as well as discussing the Critical Assessment of Techniques for Protein Structure Prediction (CASP) and the Continuous Automated Model EvaluatiOn (CAMEO) projects, and their impact on developments in the field. Furthermore, we discuss the importance of protein function prediction methods for tackling 21st century problems.