Graph Convolutional Neural Networks for Predicting Drug-Target Interactions

GraphEGFR: Multi-task and transfer learning based on molecular graph attention mechanism and fingerprints improving inhibitor bioactivity prediction for EGFR family proteins on data scarcity

The proteins within the human epidermal growth factor receptor (EGFR) family, members of the tyrosine kinase receptor family, play a pivotal role in the molecular mechanisms driving the development of various tumors. Tyrosine kinase inhibitors, key compounds in targeted therapy, encounter challenges in cancer treatment due to emerging drug resistance mutations. Consequently, machine learning has undergone significant evolution to address the challenges of cancer drug discovery related to EGFR family proteins. However, the application of deep learning in this area is hindered by inherent difficulties associated with small-scale data, particularly the risk of overfitting. Moreover, the design of a model architecture that facilitates learning through multi-task and transfer learning, coupled with appropriate molecular representation, poses substantial challenges. In this study, we introduce GraphEGFR, a deep learning regression model designed to enhance molecular representation and model architecture for predicting the bioactivity of inhibitors against both wild-type and mutant EGFR family proteins. GraphEGFR integrates a graph attention mechanism for molecular graphs with deep and convolutional neural networks for molecular fingerprints. We observed that GraphEGFR models employing multi-task and transfer learning strategies generally achieve predictive performance comparable to existing competitive methods. The integration of molecular graphs and fingerprints adeptly captures relationships between atoms and enables both global and local pattern recognition. We further validated potential multi-targeted inhibitors for wild-type and mutant HER1 kinases, exploring key amino acid residues through molecular dynamics simulations to understand molecular interactions. This predictive model offers a robust strategy that could significantly contribute to overcoming the challenges of developing deep learning models for drug discovery with limited data and exploring new frontiers in multi-targeted kinase drug discovery for EGFR family proteins.

RhoMax: Computational Prediction of Rhodopsin Absorption Maxima Using Geometric Deep Learning

Microbial rhodopsins (MRs) are a diverse and abundant family of photoactive membrane proteins that serve as model systems for biophysical techniques. Optogenetics utilizes genetic engineering to insert specialized proteins into specific neurons or brain regions, allowing for manipulation of their activity through light and enabling the mapping and control of specific brain areas in living organisms. The obstacle of optogenetics lies in the fact that light has a limited ability to penetrate biological tissues, particularly blue light in the visible spectrum. Despite this challenge, most optogenetic systems rely on blue light due to the scarcity of red-shifted opsins. Finding additional red-shifted rhodopsins would represent a major breakthrough in overcoming the challenge of limited light penetration in optogenetics. However, determining the wavelength absorption maxima for rhodopsins based on their protein sequence is a significant hurdle. Current experimental methods are time-consuming, while computational methods lack accuracy. The paper introduces a new computational approach called RhoMax that utilizes structure-based geometric deep learning to predict the absorption wavelength of rhodopsins solely based on their sequences. The method takes advantage of AlphaFold2 for accurate modeling of rhodopsin structures. Once trained on a balanced train set, RhoMax rapidly and precisely predicted the maximum absorption wavelength of more than half of the sequences in our test set with an accuracy of 0.03 eV. By leveraging computational methods for absorption maxima determination, we can drastically reduce the time needed for designing new red-shifted microbial rhodopsins, thereby facilitating advances in the field of optogenetics.

Potential inhibitors of VEGFR1, VEGFR2, and VEGFR3 developed through Deep Learning for the treatment of Cervical Cancer

Cervical cancer stands as a prevalent gynaecologic malignancy affecting women globally, often linked to persistent human papillomavirus infection. Biomarkers associated with cervical cancer, including VEGF-A, VEGF-B, VEGF-C, VEGF-D, and VEGF-E, show upregulation and are linked to angiogenesis and lymphangiogenesis. This research aims to employ in-silico methods to target tyrosine kinase receptor proteins-VEGFR-1, VEGFR-2, and VEGFR-3, and identify novel inhibitors for Vascular Endothelial Growth Factors receptors (VEGFRs). A comprehensive literary study was conducted which identified 26 established inhibitors for VEGFR-1, VEGFR-2, and VEGFR-3 receptor proteins. Compounds with high-affinity scores, including PubChem ID-25102847, 369976, and 208908 were chosen from pre-existing compounds for creating Deep Learning-based models. RD-Kit, a Deep learning algorithm, was used to generate 43 million compounds for VEGFR-1, VEGFR-2, and VEGFR-3 targets. Molecular docking studies were conducted on the top 10 molecules for each target to validate the receptor-ligand binding affinity. The results of Molecular Docking indicated that PubChem IDs-71465,645 and 11152946 exhibited strong affinity, designating them as the most efficient molecules. To further investigate their potential, a Molecular Dynamics Simulation was performed to assess conformational stability, and a pharmacophore analysis was also conducted for indoctrinating interactions.

An end-to-end method for predicting compound-protein interactions based on simplified homogeneous graph convolutional network and pre-trained language model

Identification of interactions between chemical compounds and proteins is crucial for various applications, including drug discovery, target identification, network pharmacology, and elucidation of protein functions. Deep neural network-based approaches are becoming increasingly popular in efficiently identifying compound-protein interactions with high-throughput capabilities, narrowing down the scope of candidates for traditional labor-intensive, time-consuming and expensive experimental techniques. In this study, we proposed an end-to-end approach termed SPVec-SGCN-CPI, which utilized simplified graph convolutional network (SGCN) model with low-dimensional and continuous features generated from our previously developed model SPVec and graph topology information to predict compound-protein interactions. The SGCN technique, dividing the local neighborhood aggregation and nonlinearity layer-wise propagation steps, effectively aggregates K-order neighbor information while avoiding neighbor explosion and expediting training. The performance of the SPVec-SGCN-CPI method was assessed across three datasets and compared against four machine learning- and deep learning-based methods, as well as six state-of-the-art methods. Experimental results revealed that SPVec-SGCN-CPI outperformed all these competing methods, particularly excelling in unbalanced data scenarios. By propagating node features and topological information to the feature space, SPVec-SGCN-CPI effectively incorporates interactions between compounds and proteins, enabling the fusion of heterogeneity. Furthermore, our method scored all unlabeled data in ChEMBL, confirming the top five ranked compound-protein interactions through molecular docking and existing evidence. These findings suggest that our model can reliably uncover compound-protein interactions within unlabeled compound-protein pairs, carrying substantial implications for drug re-profiling and discovery. In summary, SPVec-SGCN demonstrates its efficacy in accurately predicting compound-protein interactions, showcasing potential to enhance target identification and streamline drug discovery processes.Scientific contributionsThe methodology presented in this work not only enables the comparatively accurate prediction of compound-protein interactions but also, for the first time, take sample imbalance which is very common in real world and computation efficiency into consideration simultaneously, accelerating the target identification and drug discovery process.

Hierarchical multimodal self-attention-based graph neural network for DTI prediction

Drug-target interactions (DTIs) are a key part of drug development process and their accurate and efficient prediction can significantly boost development efficiency and reduce development time. Recent years have witnessed the rapid advancement of deep learning, resulting in an abundance of deep learning-based models for DTI prediction. However, most of these models used a single representation of drugs and proteins, making it difficult to comprehensively represent their characteristics. Multimodal data fusion can effectively compensate for the limitations of single-modal data. However, existing multimodal models for DTI prediction do not take into account both intra- and inter-modal interactions simultaneously, resulting in limited presentation capabilities of fused features and a reduction in DTI prediction accuracy. A hierarchical multimodal self-attention-based graph neural network for DTI prediction, called HMSA-DTI, is proposed to address multimodal feature fusion. Our proposed HMSA-DTI takes drug SMILES, drug molecular graphs, protein sequences and protein 2-mer sequences as inputs, and utilizes a hierarchical multimodal self-attention mechanism to achieve deep fusion of multimodal features of drugs and proteins, enabling the capture of intra- and inter-modal interactions between drugs and proteins. It is demonstrated that our proposed HMSA-DTI has significant advantages over other baseline methods on multiple evaluation metrics across five benchmark datasets.

Prediction of drug-protein interaction based on dual channel neural networks with attention mechanism

The precise identification of drug-protein inter action (DPI) can significantly speed up the drug discovery process. Bioassay methods are time-consuming and expensive to screen for each pair of drug proteins. Machine-learning-based methods cannot accurately predict a large number of DPIs. Compared with traditional computing methods, deep learning methods need less domain knowledge and have strong data learning ability. In this study, we construct a DPI prediction model based on dual channel neural networks with an efficient path attention mechanism, called DCA-DPI. The drug molecular graph and protein sequence are used as the data input of the model, and the residual graph neural network and the residual convolution network are used to learn the feature representation of the drug and protein, respectively, to obtain the feature vector of the drug and the hidden vector of protein. To get a more accurate protein feature vector, the weighted sum of the hidden vector of protein is applied using the neural attention mechanism. In the end, drug and protein vectors are concatenated and input into the full connection layer for classification. In order to evaluate the performance of DCA-DPI, three widely used public data, Human, C.elegans and DUD-E, are used in the experiment. The evaluation metrics values in the experiment are superior to other relevant methods. Experiments show that our model is efficient for DPI prediction.

G-K BertDTA: A graph representation learning and semantic embedding-based framework for drug-target affinity prediction

Developing new drugs is costly, time-consuming, and risky. Drug-target affinity (DTA), indicating the binding capability between drugs and target proteins, is a crucial indicator for drug development. Accurately predicting interaction strength between new drug-target pairs by analyzing previous experiments aids in screening potential drug molecules, repurposing them, and developing safe and effective medicines. Existing computational models for DTA prediction rely on strings or single-graph neural networks, lacking consideration of protein structure and molecular semantic information, leading to limited accuracy. Our experiments demonstrate that string-based methods may overlook protein conformations, causing a high root mean square error (RMSE) of 3.584 in affinity due to a lack of spatial context. Single graph networks also underperform on topology features, with a 6% lower confidence interval (CI) for activity classification. Absent semantic information also limits generalization across diverse compounds, resulting in 18% increment in RMSE and 5% in misclassifications within quantifications study, restricting potential drug discovery. To address these limitations, we propose G-K BertDTA, a novel framework for accurate DTA prediction incorporating protein features, molecular semantic features, and molecular structural information. In this proposed model, we represent drugs as graphs, with a GIN employed to learn the molecular topological information. For the extraction of protein structural features, we utilize a DenseNet architecture. A knowledge-based BERT semantic model is incorporated to obtain rich pre-trained semantic embeddings, thereby enhancing the feature information. We extensively evaluated our proposed approach on the publicly available benchmark datasets (i.e., KIBA and Davis), and experimental results demonstrate the promising performance of our method, which consistently outperforms previous state-of-the-art approaches. Code is available at https://github.com/AmbitYuki/G-K-BertDTA.

Advances in machine intelligence-driven virtual screening approaches for big-data

Virtual screening (VS) is an integral and ever-evolving domain of drug discovery framework. The VS is traditionally classified into ligand-based (LB) and structure-based (SB) approaches. Machine intelligence or artificial intelligence has wide applications in the drug discovery domain to reduce time and resource consumption. In combination with machine intelligence algorithms, VS has emerged into revolutionarily progressive technology that learns within robust decision orders for data curation and hit molecule screening from large VS libraries in minutes or hours. The exponential growth of chemical and biological data has evolved as "big-data" in the public domain demands modern and advanced machine intelligence-driven VS approaches to screen hit molecules from ultra-large VS libraries. VS has evolved from an individual approach (LB and SB) to integrated LB and SB techniques to explore various ligand and target protein aspects for the enhanced rate of appropriate hit molecule prediction. Current trends demand advanced and intelligent solutions to handle enormous data in drug discovery domain for screening and optimizing hits or lead with fewer or no false positive hits. Following the big-data drift and tremendous growth in computational architecture, we presented this review. Here, the article categorized and emphasized individual VS techniques, detailed literature presented for machine learning implementation, modern machine intelligence approaches, and limitations and deliberated the future prospects.

Drug-Online: an online platform for drug-target interaction, affinity, and binding sites identification using deep learning

BACKGROUND

Accurately identifying drug-target interaction (DTI), affinity (DTA), and binding sites (DTS) is crucial for drug screening, repositioning, and design, as well as for understanding the functions of target. Although there are a few online platforms based on deep learning for drug-target interaction, affinity, and binding sites identification, there is currently no integrated online platforms for all three aspects.

RESULTS

Our solution, the novel integrated online platform Drug-Online, has been developed to facilitate drug screening, target identification, and understanding the functions of target in a progressive manner of "interaction-affinity-binding sites". Drug-Online platform consists of three parts: the first part uses the drug-target interaction identification method MGraphDTA, based on graph neural networks (GNN) and convolutional neural networks (CNN), to identify whether there is a drug-target interaction. If an interaction is identified, the second part employs the drug-target affinity identification method MMDTA, also based on GNN and CNN, to calculate the strength of drug-target interaction, i.e., affinity. Finally, the third part identifies drug-target binding sites, i.e., pockets. The method pt-lm-gnn used in this part is also based on GNN.

CONCLUSIONS

Drug-Online is a reliable online platform that integrates drug-target interaction, affinity, and binding sites identification. It is freely available via the Internet at http://39.106.7.26:8000/Drug-Online/ .

HyperPCM: Robust Task-Conditioned Modeling of Drug-Target Interactions

A central problem in drug discovery is to identify the interactions between drug-like compounds and protein targets. Over the past few decades, various quantitative structure-activity relationship (QSAR) and proteo-chemometric (PCM) approaches have been developed to model and predict these interactions. While QSAR approaches solely utilize representations of the drug compound, PCM methods incorporate both representations of the protein target and the drug compound, enabling them to achieve above-chance predictive accuracy on previously unseen protein targets. Both QSAR and PCM approaches have recently been improved by machine learning and deep neural networks, that allow the development of drug-target interaction prediction models from measurement data. However, deep neural networks typically require large amounts of training data and cannot robustly adapt to new tasks, such as predicting interaction for unseen protein targets at inference time. In this work, we propose to use HyperNetworks to efficiently transfer information between tasks during inference and thus to accurately predict drug-target interactions on unseen protein targets. Our HyperPCM method reaches state-of-the-art performance compared to previous methods on multiple well-known benchmarks, including Davis, DUD-E, and a ChEMBL derived data set, and particularly excels at zero-shot inference involving unseen protein targets. Our method, as well as reproducible data preparation, is available at https://github.com/ml-jku/hyper-dti.

Explainable Graph Neural Networks with Data Augmentation for Predicting pKa of C-H Acids

The pKa of C-H acids is an important parameter in the fields of organic synthesis, drug discovery, and materials science. However, the prediction of pKa is still a great challenge due to the limit of experimental data and the lack of chemical insight. Here, a new model for predicting the pKa values of C-H acids is proposed on the basis of graph neural networks (GNNs) and data augmentation. A message passing unit (MPU) was used to extract the topological and target-related information from the molecular graph data, and a readout layer was utilized to retrieve the information on the ionization site C atom. The retrieved information then was adopted to predict pKa by a fully connected network. Furthermore, to increase the diversity of the training data, a knowledge-infused data augmentation technique was established by replacing the H atoms in a molecule with substituents exhibiting different electronic effects. The MPU was pretrained with the augmented data. The efficacy of data augmentation was confirmed by visualizing the distribution of compounds with different substituents and by classifying compounds. The explainability of the model was studied by examining the change of pKa values when a specific atom was masked. This explainability was used to identify the key substituents for pKa. The model was evaluated on two data sets from the iBonD database. Dataset1 includes the experimental pKa values of C-H acids measured in DMSO, while dataset2 comprises the pKa values measured in water. The results show that the knowledge-infused data augmentation technique greatly improves the predictive accuracy of the model, especially when the number of samples is small.

PLANET: A Multi-objective Graph Neural Network Model for Protein-Ligand Binding Affinity Prediction

Predicting protein-ligand binding affinity is a central issue in drug design. Various deep learning models have been published in recent years, where many of them rely on 3D protein-ligand complex structures as input and tend to focus on the single task of reproducing binding affinity. In this study, we have developed a graph neural network model called PLANET (Protein-Ligand Affinity prediction NETwork). This model takes the graph-represented 3D structure of the binding pocket on the target protein and the 2D chemical structure of the ligand molecule as input. It was trained through a multi-objective process with three related tasks, including deriving the protein-ligand binding affinity, protein-ligand contact map, and ligand distance matrix. Besides the protein-ligand complexes with known binding affinity data retrieved from the PDBbind database, a large number of non-binder decoys were also added to the training data for deriving the final model of PLANET. When tested on the CASF-2016 benchmark, PLANET exhibited a scoring power comparable to the best result yielded by other deep learning models as well as a reasonable ranking power and docking power. In virtual screening trials conducted on the DUD-E benchmark, PLANET's performance was notably better than several deep learning and machine learning models. As on the LIT-PCBA benchmark, PLANET achieved comparable accuracy as the conventional docking program Glide, but it only spent less than 1% of Glide's computation time to finish the same job because PLANET did not need exhaustive conformational sampling. Considering the decent accuracy and efficiency of PLANET in binding affinity prediction, it may become a useful tool for conducting large-scale virtual screening.

PfgPDI: Pocket feature-enabled graph neural network for protein-drug interaction prediction

Biomolecular interaction recognition between ligands and proteins is an essential task, which largely enhances the safety and efficacy in drug discovery and development stage. Studying the interaction between proteins and ligands can improve the understanding of disease pathogenesis and lead to more effective drug targets. Additionally, it can aid in determining drug parameters, ensuring proper absorption, distribution, and metabolism within the body. Due to incomplete feature representation or the model's inadequate adaptation to protein-ligand complexes, the existing methodologies suffer from suboptimal predictive accuracy. To address these pitfalls, in this study, we designed a new deep learning method based on transformer and GCN. We first utilized the transformer network to grasp crucial information of the original protein sequences within the smile sequences and connected them to prevent falling into a local optimum. Furthermore, a series of dilation convolutions are performed to obtain the pocket features and smile features, subsequently subjected to graphical convolution to optimize the connections. The combined representations are fed into the proposed model for classification prediction. Experiments conducted on various protein-ligand binding prediction methods prove the effectiveness of our proposed method. It is expected that the PfgPDI can contribute to drug prediction and accelerate the development of new drugs, while also serving as a valuable partner for drug testing and Research and Development engineers.

A bidirectional interpretable compound-protein interaction prediction framework based on cross attention

The identification of compound-protein interactions (CPIs) plays a vital role in drug discovery. However, the huge cost and labor-intensive nature in vitro and vivo experiments make it urgent for researchers to develop novel CPI prediction methods. Despite emerging deep learning methods have achieved promising performance in CPI prediction, they also face ongoing challenges: (i) providing bidirectional interpretability from both the chemical and biological perspective for the prediction results; (ii) comprehensively evaluating model generalization performance; (iii) demonstrating the practical applicability of these models. To overcome the challenges posed by current deep learning methods, we propose a cross multi-head attention oriented bidirectional interpretable CPI prediction model (CmhAttCPI). First, CmhAttCPI takes molecular graphs and protein sequences as inputs, utilizing the GCW module to learn atom features and the CNN module to learn residue features, respectively. Second, the model applies cross multi-head attention module to compute attention weights for atoms and residues. Finally, CmhAttCPI employs a fully connected neural network to predict scores for CPIs. We evaluated the performance of CmhAttCPI on balanced datasets and imbalanced datasets. The results consistently show that CmhAttCPI outperforms multiple state-of-the-art methods. We constructed three scenarios based on compound and protein clustering and comprehensively evaluated the model generalization ability within these scenarios. The results demonstrate that the generalization ability of CmhAttCPI surpasses that of other models. Besides, the visualizations of attention weights reveal that CmhAttCPI provides chemical and biological interpretation for CPI prediction. Moreover, case studies confirm the practical applicability of CmhAttCPI in discovering anticancer candidates.

Directional ΔG Neural Network (DrΔG-Net): A Modular Neural Network Approach to Binding Free Energy Prediction

The protein-ligand binding free energy is a central quantity in structure-based computational drug discovery efforts. Although popular alchemical methods provide sound statistical means of computing the binding free energy of a large breadth of systems, they are generally too costly to be applied at the same frequency as end point or ligand-based methods. By contrast, these data-driven approaches are typically fast enough to address thousands of systems but with reduced transferability to unseen systems. We introduce DrΔG-Net (or simply Dragnet), an equivariant graph neural network that can blend ligand-based and protein-ligand data-driven approaches. It is based on a 3D fingerprint representation of the ligand alone and in complex with the protein target. Dragnet is a global scoring function to predict the binding affinity of arbitrary protein-ligand complexes, but can be easily tuned via transfer learning to specific systems or end points, performing similarly to common 2D ligand-based approaches in these tasks. Dragnet is evaluated on a total of 28 validation proteins with a set of congeneric ligands derived from the Binding DB and one custom set extracted from the ChEMBL Database. In general, a handful of experimental binding affinities are sufficient to optimize the scoring function for a particular protein and ligand scaffold. When not available, predictions from physics-based methods such as absolute free energy perturbation can be used for the transfer learning tuning of Dragnet. Furthermore, we use our data to illustrate the present limitations of data-driven modeling of binding free energy predictions.

Drug-drug interactions prediction based on deep learning and knowledge graph: A review

Drug-drug interactions (DDIs) can produce unpredictable pharmacological effects and lead to adverse events that have the potential to cause irreversible damage to the organism. Traditional methods to detect DDIs through biological or pharmacological analysis are time-consuming and expensive, therefore, there is an urgent need to develop computational methods to effectively predict drug-drug interactions. Currently, deep learning and knowledge graph techniques which can effectively extract features of entities have been widely utilized to develop DDI prediction methods. In this research, we aim to systematically review DDI prediction researches applying deep learning and graph knowledge. The available biomedical data and public databases related to drugs are firstly summarized in this review. Then, we discuss the existing drug-drug interactions prediction methods which have utilized deep learning and knowledge graph techniques and group them into three main classes: deep learning-based methods, knowledge graph-based methods, and methods that combine deep learning with knowledge graph. We comprehensively analyze the commonly used drug related data and various DDI prediction methods, and compare these prediction methods on benchmark datasets. Finally, we briefly discuss the challenges related to drug-drug interactions prediction, including asymmetric DDIs prediction and high-order DDI prediction.

GraphsformerCPI: Graph Transformer for Compound-Protein Interaction Prediction

Accurately predicting compound-protein interactions (CPI) is a critical task in computer-aided drug design. In recent years, the exponential growth of compound activity and biomedical data has highlighted the need for efficient and interpretable prediction approaches. In this study, we propose GraphsformerCPI, an end-to-end deep learning framework that improves prediction performance and interpretability. GraphsformerCPI treats compounds and proteins as sequences of nodes with spatial structures, and leverages novel structure-enhanced self-attention mechanisms to integrate semantic and graph structural features within molecules for deep molecule representations. To capture the vital association between compound atoms and protein residues, we devise a dual-attention mechanism to effectively extract relational features through .cross-mapping. By extending the powerful learning capabilities of Transformers to spatial structures and extensively utilizing attention mechanisms, our model offers strong interpretability, a significant advantage over most black-box deep learning methods. To evaluate GraphsformerCPI, extensive experiments were conducted on benchmark datasets including human, C. elegans, Davis and KIBA datasets. We explored the impact of model depth and dropout rate on performance and compared our model against state-of-the-art baseline models. Our results demonstrate that GraphsformerCPI outperforms baseline models in classification datasets and achieves competitive performance in regression datasets. Specifically, on the human dataset, GraphsformerCPI achieves an average improvement of 1.6% in AUC, 0.5% in precision, and 5.3% in recall. On the KIBA dataset, the average improvement in Concordance index (CI) and mean squared error (MSE) is 3.3% and 7.2%, respectively. Molecular docking shows that our model provides novel insights into the intrinsic interactions and binding mechanisms. Our research holds practical significance in effectively predicting CPIs and binding affinities, identifying key atoms and residues, enhancing model interpretability.

Drugging the entire human proteome: Are we there yet?

Each of the ∼20,000 proteins in the human proteome is a potential target for compounds that bind to it and modify its function. The 3D structures of most of these proteins are now available. Here, we discuss the prospects for using these structures to perform proteome-wide virtual HTS (VHTS). We compare physics-based (docking) and AI VHTS approaches, some of which are now being applied with large databases of compounds to thousands of targets. Although preliminary proteome-wide screens are now within our grasp, further methodological developments are expected to improve the accuracy of the results.

TransVAE-DTA: Transformer and variational autoencoder network for drug-target binding affinity prediction

BACKGROUND AND OBJECTIVE

Recent studies have emphasized the significance of computational in silico drug-target binding affinity (DTA) prediction in the field of drug discovery and drug repurposing. However, existing DTA prediction approaches suffer from two major deficiencies that impede their progress. Firstly, while most methods primarily focus on the feature representations of drug-target binding affinity pairs, they fail to consider the long-distance relationships of proteins. Furthermore, many deep learning-based DTA predictors simply model the interaction of drug-target pairs through concatenation, which hampers the ability to enhance prediction performance.

METHODS

To address these issues, this study proposes a novel framework named TransVAE-DTA, which combines the transformer and variational autoencoder (VAE). Inspired by the early success of VAEs, we aim to further investigate the feasibility of VAEs for drug structure encoding, while utilizing the transformer architecture for target feature representation. Additionally, an adaptive attention pooling (AAP) module is designed to fuse the drug and target encoded features. Notably, TransVAE-DTA is proven to maximize the lower bound of the joint likelihood of drug, target, and their DTAs.

RESULTS

Experimental results demonstrate the superiority of TransVAE-DTA in drug-target binding affinity prediction assignments on two public Davis and KIBA datasets.

CONCLUSIONS

In this research, the developed TransVAE-DTA opens a new avenue for engineering drug-target interactions.

GATLGEMF: A graph attention model with line graph embedding multi-complex features for ncRNA-protein interactions prediction

Non-coding RNA (ncRNA) plays an important role in many fundamental biological processes, and it may be closely associated with many complex human diseases. NcRNAs exert their functions by interacting with proteins. Therefore, identifying novel ncRNA-protein interactions (NPIs) is important for understanding the mechanism of ncRNAs role. The computational approach has the advantage of low cost and high efficiency. Machine learning and deep learning have achieved great success by making full use of sequence information and structure information. Graph neural network (GNN) is a deep learning algorithm for complex network link prediction, which can extract and discover features in graph topology data. In this study, we propose a new computational model called GATLGEMF. We used a line graph transformation strategy to obtain the most valuable feature information and input this feature information into the attention network to predict NPIs. The results on four benchmark datasets show that our method achieves superior performance. We further compare GATLGEMF with the state-of-the-art existing methods to evaluate the model performance. GATLGEMF shows the best performance with the area under curve (AUC) of 92.41% and 98.93% on RPI2241 and NPInter v2.0 datasets, respectively. In addition, a case study shows that GATLGEMF has the ability to predict new interactions based on known interactions. The source code is available at https://github.com/JianjunTan-Beijing/GATLGEMF.

Exploring protein-ligand binding affinity prediction with electron density-based geometric deep learning

Rational structure-based drug design relies on accurate predictions of protein-ligand binding affinity from structural molecular information. Although deep learning-based methods for predicting binding affinity have shown promise in computational drug design, certain approaches have faced criticism for their potential to inadequately capture the fundamental physical interactions between ligands and their macromolecular targets or for being susceptible to dataset biases. Herein, we propose to include bond-critical points based on the electron density of a protein-ligand complex as a fundamental physical representation of protein-ligand interactions. Employing a geometric deep learning model, we explore the usefulness of these bond-critical points to predict absolute binding affinities of protein-ligand complexes, benchmark model performance against existing methods, and provide a critical analysis of this new approach. The models achieved root-mean-squared errors of 1.4-1.8 log units on the PDBbind dataset, and 1.0-1.7 log units on the PDE10A dataset, not indicating significant advantages over benchmark methods, and thus rendering the utility of electron density for deep learning models context-dependent. The relationship between intermolecular electron density and corresponding binding affinity was analyzed, and Pearson correlation coefficients r > 0.7 were obtained for several macromolecular targets.

Artificial Intelligence for Drug Discovery: Are We There Yet?

Drug discovery is adapting to novel technologies such as data science, informatics, and artificial intelligence (AI) to accelerate effective treatment development while reducing costs and animal experiments. AI is transforming drug discovery, as indicated by increasing interest from investors, industrial and academic scientists, and legislators. Successful drug discovery requires optimizing properties related to pharmacodynamics, pharmacokinetics, and clinical outcomes. This review discusses the use of AI in the three pillars of drug discovery: diseases, targets, and therapeutic modalities, with a focus on small-molecule drugs. AI technologies, such as generative chemistry, machine learning, and multiproperty optimization, have enabled several compounds to enter clinical trials. The scientific community must carefully vet known information to address the reproducibility crisis. The full potential of AI in drug discovery can only be realized with sufficient ground truth and appropriate human intervention at later pipeline stages.

FOTF-CPI: A compound-protein interaction prediction transformer based on the fusion of optimal transport fragments

Compound-protein interaction (CPI) affinity prediction plays an important role in reducing the cost and time of drug discovery. However, the interpretability of how fragments function in CPI is impacted by the fact that current methods ignore the affinity relationships between fragments of compounds and fragments of proteins in CPI modeling. This article introduces an improved Transformer called FOTF-CPI (a Fusion of Optimal Transport Fragments compound-protein interaction prediction model). We use an optimal transport-based fragmentation approach to improve the model's understanding of compound and protein sequences. Additionally, a fused attention mechanism is employed, which combines the features of fragments to capture full affinity information. This fused attention redistributes higher attention scores to fragments with higher affinity. Experimental results show FOTF-CPI achieves an average 2% higher performance than other models on all three datasets. Furthermore, the visualization confirms the potential of FOTF-CPI for drug discovery applications.

Functional microRNA-targeting drug discovery by graph-based deep learning

MicroRNAs are recognized as key drivers in many cancers but targeting them with small molecules remains a challenge. We present RiboStrike, a deep-learning framework that identifies small molecules against specific microRNAs. To demonstrate its capabilities, we applied it to microRNA-21 (miR-21), a known driver of breast cancer. To ensure selectivity toward miR-21, we performed counter-screens against miR-122 and DICER. Auxiliary models were used to evaluate toxicity and rank the candidates. Learning from various datasets, we screened a pool of nine million molecules and identified eight, three of which showed anti-miR-21 activity in both reporter assays and RNA sequencing experiments. Target selectivity of these compounds was assessed using microRNA profiling and RNA sequencing analysis. The top candidate was tested in a xenograft mouse model of breast cancer metastasis, demonstrating a significant reduction in lung metastases. These results demonstrate RiboStrike's ability to nominate compounds that target the activity of miRNAs in cancer.

AttentionMGT-DTA: A multi-modal drug-target affinity prediction using graph transformer and attention mechanism

The accurate prediction of drug-target affinity (DTA) is a crucial step in drug discovery and design. Traditional experiments are very expensive and time-consuming. Recently, deep learning methods have achieved notable performance improvements in DTA prediction. However, one challenge for deep learning-based models is appropriate and accurate representations of drugs and targets, especially the lack of effective exploration of target representations. Another challenge is how to comprehensively capture the interaction information between different instances, which is also important for predicting DTA. In this study, we propose AttentionMGT-DTA, a multi-modal attention-based model for DTA prediction. AttentionMGT-DTA represents drugs and targets by a molecular graph and binding pocket graph, respectively. Two attention mechanisms are adopted to integrate and interact information between different protein modalities and drug-target pairs. The experimental results showed that our proposed model outperformed state-of-the-art baselines on two benchmark datasets. In addition, AttentionMGT-DTA also had high interpretability by modeling the interaction strength between drug atoms and protein residues. Our code is available at https://github.com/JK-Liu7/AttentionMGT-DTA.

Comprehensive Review on Drug-target Interaction Prediction - Latest Developments and Overview

Drug-target interactions (DTIs) are an important part of the drug development process. When the drug (a chemical molecule) binds to a target (proteins or nucleic acids), it modulates the biological behavior/function of the target, returning it to its normal state. Predicting DTIs plays a vital role in the drug discovery (DD) process as it has the potential to enhance efficiency and reduce costs. However, DTI prediction poses significant challenges and expenses due to the time-consuming and costly nature of experimental assays. As a result, researchers have increased their efforts to identify the association between medications and targets in the hopes of speeding up drug development and shortening the time to market. This paper provides a detailed discussion of the initial stage in drug discovery, namely drug-target interactions. It focuses on exploring the application of machine learning methods within this step. Additionally, we aim to conduct a comprehensive review of relevant papers and databases utilized in this field. Drug target interaction prediction covers a wide range of applications: drug discovery, prediction of adverse effects and drug repositioning. The prediction of drugtarget interactions can be categorized into three main computational methods: docking simulation approaches, ligand-based methods, and machine-learning techniques.

Binding Thermodynamics of Fourth-Generation EGFR Inhibitors Revealed by Absolute Binding Free Energy Calculations

The overexpression or mutation of the kinase domain of the epidermal growth factor receptor (EGFR) is strongly associated with non-small-cell lung cancer (NSCLC). EGFR tyrosine kinase inhibitors (TKIs) have proven to be effective in treating NSCLC patients. However, EGFR mutations can result in drug resistance. To elucidate the mechanisms underlying this resistance and inform future drug development, we examined the binding affinities of BLU-945, a recently reported fourth-generation TKI, to wild-type EGFR (EGFRWT) and its double-mutant (L858R/T790M; EGFRDM) and triple-mutant (L858R/T790M/C797S; EGFRTM) forms. We compared the binding affinities of BLU-945, BLU-945 analogues, CH7233163 (another fourth-generation TKI), and erlotinib (a first-generation TKI) using absolute binding free energy calculations. Our findings reveal that BLU-945 and CH7233163 exhibit binding affinities to both EGFRDM and EGFRTM stronger than those of erlotinib, corroborating experimental data. We identified K745 and T854 as the key residues in the binding of fourth-generation EGFR TKIs. Electrostatic forces were the predominant driving force for the binding of fourth-generation TKIs to EGFR mutants. Furthermore, we discovered that the incorporation of piperidinol and sulfone groups in BLU-945 substantially enhanced its binding capacity to EGFR mutants. Our study offers valuable theoretical insights for optimizing fourth-generation EGFR TKIs.

Graph neural networks for the identification of novel inhibitors of a small RNA

MicroRNAs (miRNAs) play a crucial role in post-transcriptional gene regulation and have been implicated in various diseases, including cancers and lung disease. In recent years, Graph Neural Networks (GNNs) have emerged as powerful tools for analyzing graph-structured data, making them well-suited for the analysis of molecular structures. In this work, we explore the application of GNNs in ligand-based drug screening for small molecules targeting miR-21. By representing a known dataset of small molecules targeting miR-21 as graphs, GNNs can learn complex relationships between their structures and activities, enabling the prediction of potential miRNA-targeting small molecules by capturing the structural features and similarity between known miRNA-targeting compounds. The use of GNNs in miRNA-targeting drug screening holds promise for the discovery of novel therapeutic agents and provides a computational framework for efficient screening of large chemical libraries.

Prediction of protein-ligand binding affinity with deep learning

The prediction of binding affinities between target proteins and small molecule drugs is essential for speeding up the drug research and design process. To attain precise and effective affinity prediction, computer-aided methods are employed in the drug discovery pipeline. In the last decade, a variety of computational methods has been developed, with deep learning being the most commonly used approach. We have gathered several deep learning methods and classified them into convolutional neural networks (CNNs), graph neural networks (GNNs), and Transformers for analysis and discussion. Initially, we conducted an analysis of the different deep learning methods, focusing on their feature construction and model architecture. We discussed the advantages and disadvantages of each model. Subsequently, we conducted experiments using four deep learning methods on the PDBbind v.2016 core set. We evaluated their prediction capabilities in various affinity intervals and statistically and visually analyzed the samples of correct and incorrect predictions for each model. Through visual analysis, we attempted to combine the strengths of the four models to improve the Root Mean Square Error (RMSE) of predicted affinities by 1.6% (reducing the absolute value to 1.101) and the Pearson Correlation Coefficient (R) by 2.9% (increasing the absolute value to 0.894) compared to the current state-of-the-art method. Lastly, we discussed the challenges faced by current deep learning methods in affinity prediction and proposed potential solutions to address these issues.

Deep Learning for Drug Development: Using CNNs in MIA-QSAR to Predict Plasma Protein Binding of Drugs

Predicting plasma protein binding (PPB) is crucial in drug development due to its profound impact on drug efficacy and safety. In our study, we employed a convolutional neural network (CNN) as a tool to extract valuable information from the molecular structures of 100 different drugs. These extracted features were then used as inputs for a feedforward network to predict the PPB of each drug. Through this approach, we successfully obtained 10 specific numerical features from each drug's molecular structure, which represent fundamental aspects of their molecular composition. Leveraging the CNN's ability to capture these features significantly improved the precision of our predictions. Our modeling results revealed impressive accuracy, with an R2 train value of 0.89 for the training dataset, a [Formula: see text] of 0.98, a [Formula: see text] of 0.931 for the external validation dataset, and a low cross-validation mean squared error (CV-MSE) of 0.0213. These metrics highlight the effectiveness of our deep learning techniques in the fields of pharmacokinetics and drug development. This study makes a substantial contribution to the expanding body of research exploring the application of artificial intelligence (AI) and machine learning in drug development. By adeptly capturing and utilizing molecular features, our method holds promise for enhancing drug efficacy and safety assessments in pharmaceutical research. These findings underscore the potential for future investigations in this exciting and transformative field.

Bridging the Worlds of Pharmacometrics and Machine Learning

Precision medicine requires individualized modeling of disease and drug dynamics, with machine learning-based computational techniques gaining increasing popularity. The complexity of either field, however, makes current pharmacological problems opaque to machine learning practitioners, and state-of-the-art machine learning methods inaccessible to pharmacometricians. To help bridge the two worlds, we provide an introduction to current problems and techniques in pharmacometrics that ranges from pharmacokinetic and pharmacodynamic modeling to pharmacometric simulations, model-informed precision dosing, and systems pharmacology, and review some of the machine learning approaches to address them. We hope this would facilitate collaboration between experts, with complementary strengths of principled pharmacometric modeling and flexibility of machine learning leading to synergistic effects in pharmacological applications.

DDI-GCN: Drug-drug interaction prediction via explainable graph convolutional networks

Drug-drug interactions (DDI) may lead to unexpected side effects, which is a growing concern in both academia and industry. Many DDIs have been reported, but the underlying mechanisms are not well understood. Predicting and understanding DDIs can help researchers to improve drug safety and protect patient health. Here, we introduce DDI-GCN, a method that utilizes graph convolutional networks (GCN) to predict DDIs based on chemical structures. We demonstrate that this method achieves state-of-the-art prediction performance on the independent hold-out set. It can also provide visualization of structural features associated with DDIs, which can help us to study the underlying mechanisms. To make it easy and accessible to use, we developed a web server for DDI-GCN, which is freely available at http://wengzq-lab.cn/ddi/.

Ligand binding affinity prediction with fusion of graph neural networks and 3D structure-based complex graph

Accurate prediction of protein-ligand binding affinity is pivotal for drug design and discovery. Here, we proposed a novel deep fusion graph neural networks framework named FGNN to learn the protein-ligand interactions from the 3D structures of protein-ligand complexes. Unlike 1D sequences for proteins or 2D graphs for ligands, the 3D graph of protein-ligand complex enables the more accurate representations of the protein-ligand interactions. Benchmark studies have shown that our fusion models FGNN can achieve more accurate prediction of binding affinity than any individual algorithm. The advantages of fusion strategies have been demonstrated in terms of expressive power of data, learning efficiency and model interpretability. Our fusion models show satisfactory performances on diverse data sets, demonstrating their generalization ability. Given the good performances in both binding affinity prediction and virtual screening, our fusion models are expected to be practically applied for drug screening and design. Our work highlights the potential of the fusion graph neural network algorithm in solving complex prediction problems in computational biology and chemistry. The fusion graph neural networks (FGNN) model is freely available in https://github.com/LinaDongXMU/FGNN.

A Hybrid Structure-Based Machine Learning Approach for Predicting Kinase Inhibition by Small Molecules

Kinases have been the focus of drug discovery programs for three decades leading to over 70 therapeutic kinase inhibitors and biophysical affinity measurements for over 130,000 kinase-compound pairs. Nonetheless, the precise target spectrum for many kinases remains only partly understood. In this study, we describe a computational approach to unlocking qualitative and quantitative kinome-wide binding measurements for structure-based machine learning. Our study has three components: (i) a Kinase Inhibitor Complex (KinCo) data set comprising in silico predicted kinase structures paired with experimental binding constants, (ii) a machine learning loss function that integrates qualitative and quantitative data for model training, and (iii) a structure-based machine learning model trained on KinCo. We show that our approach outperforms methods trained on crystal structures alone in predicting binary and quantitative kinase-compound interaction affinities; relative to structure-free methods, our approach also captures known kinase biochemistry and more successfully generalizes to distant kinase sequences and compound scaffolds.

Drug-Target Interaction Prediction via Graph Auto-Encoder and Multi-Subspace Deep Neural Networks

Computational prediction of drug-target interaction (DTI) is important for the new drug discovery. Currently, the deep neural network (DNN) has been widely used in DTI prediction. However, parameters of the DNN could be insufficiently trained and features of the data could be insufficiently utilized, because the DTI data is limited and its dimension is very high. To deal with the above problems, in this paper, a graph auto-encoder and multi-subspace deep neural network (GAEMSDNN) is designed. GAEMSDNN enhances its learning ability with a graph auto-encoder, a subspace layer and an ensemble layer. The graph auto-encoder can preserve the reconstruction information. The subspace layer can obtain different strong feature subsets. The ensemble layer in the GAEMSDNN can comprehensively utilize these strong feature subsets in a unified optimization framework. As a result, more features can be extracted from the network input and the DNN network can be better trained. In experiments, the results of GAEMSDNN are significantly improved compared to the previous methods, which validates the effectiveness of our strategies.

A deep learning framework for accurate reaction prediction and its application on high-throughput experimentation data

In recent years, it has been seen that artificial intelligence (AI) starts to bring revolutionary changes to chemical synthesis. However, the lack of suitable ways of representing chemical reactions and the scarceness of reaction data has limited the wider application of AI to reaction prediction. Here, we introduce a novel reaction representation, GraphRXN, for reaction prediction. It utilizes a universal graph-based neural network framework to encode chemical reactions by directly taking two-dimension reaction structures as inputs. The GraphRXN model was evaluated by three publically available chemical reaction datasets and gave on-par or superior results compared with other baseline models. To further evaluate the effectiveness of GraphRXN, wet-lab experiments were carried out for the purpose of generating reaction data. GraphRXN model was then built on high-throughput experimentation data and a decent accuracy (R2 of 0.712) was obtained on our in-house data. This highlights that the GraphRXN model can be deployed in an integrated workflow which combines robotics and AI technologies for forward reaction prediction.

Recent Developments in Ultralarge and Structure-Based Virtual Screening Approaches

Drug development is a wide scientific field that faces many challenges these days. Among them are extremely high development costs, long development times, and a small number of new drugs that are approved each year. New and innovative technologies are needed to solve these problems that make the drug discovery process of small molecules more time and cost efficient, and that allow previously undruggable receptor classes to be targeted, such as protein-protein interactions. Structure-based virtual screenings (SBVSs) have become a leading contender in this context. In this review, we give an introduction to the foundations of SBVSs and survey their progress in the past few years with a focus on ultralarge virtual screenings (ULVSs). We outline key principles of SBVSs, recent success stories, new screening techniques, available deep learning-based docking methods, and promising future research directions. ULVSs have an enormous potential for the development of new small-molecule drugs and are already starting to transform early-stage drug discovery.

A generalized protein-ligand scoring framework with balanced scoring, docking, ranking and screening powers

Applying machine learning algorithms to protein-ligand scoring functions has aroused widespread attention in recent years due to the high predictive accuracy and affordable computational cost. Nevertheless, most machine learning-based scoring functions are only applicable to a specific task, e.g., binding affinity prediction, binding pose prediction or virtual screening, suggesting that the development of a scoring function with balanced performance in all critical tasks remains a grand challenge. To this end, we propose a novel parameterization strategy by introducing an adjustable binding affinity term that represents the correlation between the predicted outcomes and experimental data into the training of mixture density network. The resulting residue-atom distance likelihood potential not only retains the superior docking and screening power over all the other state-of-the-art approaches, but also achieves a remarkable improvement in scoring and ranking performance. We emphatically explore the impacts of several key elements on prediction accuracy as well as the task preference, and demonstrate that the performance of scoring/ranking and docking/screening tasks of a certain model could be well balanced through an appropriate manner. Overall, our study highlights the potential utility of our innovative parameterization strategy as well as the resulting scoring framework in future structure-based drug design.

PredPS: Attention-based graph neural network for predicting stability of compounds in human plasma

Stability of compounds in the human plasma is crucial for maintaining sufficient systemic drug exposure and considered an essential factor in the early stages of drug discovery and development. The rapid degradation of compounds in the plasma can result in poor in vivo efficacy. Currently, there are no open-source software programs for predicting human plasma stability. In this study, we developed an attention-based graph neural network, PredPS to predict the plasma stability of compounds in human plasma using in-house and open-source datasets. The PredPS outperformed the two machine learning and two deep learning algorithms that were used for comparison indicating its stability-predicting efficiency. PredPS achieved an area under the receiver operating characteristic curve of 90.1%, accuracy of 83.5%, sensitivity of 82.3%, and specificity of 84.6% when evaluated using 5-fold cross-validation. In the early stages of drug discovery, PredPS could be a helpful method for predicting the human plasma stability of compounds. Saving time and money can be accomplished by adopting an in silico-based plasma stability prediction model at the high-throughput screening stage. The source code for PredPS is available at https://bitbucket.org/krict-ai/predps and the PredPS web server is available at https://predps.netlify.app.

AQDnet: Deep Neural Network for Protein-Ligand Docking Simulation

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

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

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

LigBind: identifying binding residues for over 1000 ligands with relation-aware graph neural networks

Identifying the interactions between proteins and ligands is significant for drug discovery and design. Considering the diverse binding patterns of ligands, the ligand-specific methods are trained per ligand to predict binding residues. However, most of the existing ligand-specific methods ignore shared binding preferences among various ligands and generally only cover a limited number of ligands with a sufficient number of known binding proteins. In this study, we propose a relation-aware framework LigBind with graph-level pre-training to enhance the ligand-specific binding residue predictions for 1159 ligands, which can effectively cover the ligands with a few known binding proteins. LigBind first pre-trains a graph neural network-based feature extractor for ligand-residue pairs and relation-aware classifiers for similar ligands. Then, LigBind is fine-tuned with ligand-specific binding data, where a domain adaptive neural network is designed to automatically leverage the diversity and similarity of various ligand-binding patterns for accurate binding residue prediction. We construct ligand-specific benchmark datasets of 1159 ligands and 16 unseen ligands, which are used to evaluate the effectiveness of LigBind. The results demonstrate the LigBind's efficacy on the large-scale ligand-specific benchmark datasets, and generalizes well to unseen ligands. LigBind also enables accurate identification of the ligand-binding residues in the main protease, papain-like protease and the RNA-dependent RNA polymerase of SARS-CoV-2. The webserver and source codes of LigBind are available at http://www.csbio.sjtu.edu.cn/bioinf/LigBind/ and https://github.com/YYingXia/LigBind/ for academic use.

Structure-based drug design with geometric deep learning

Structure-based drug design uses three-dimensional geometric information of macromolecules, such as proteins or nucleic acids, to identify suitable ligands. Geometric deep learning, an emerging concept of neural-network-based machine learning, has been applied to macromolecular structures. This review provides an overview of the recent applications of geometric deep learning in bioorganic and medicinal chemistry, highlighting its potential for structure-based drug discovery and design. Emphasis is placed on molecular property prediction, ligand binding site and pose prediction, and structure-based de novo molecular design. The current challenges and opportunities are highlighted, and a forecast of the future of geometric deep learning for drug discovery is presented.

Multimodal learning with graphs

Artificial intelligence for graphs has achieved remarkable success in modeling complex systems, ranging from dynamic networks in biology to interacting particle systems in physics. However, the increasingly heterogeneous graph datasets call for multimodal methods that can combine different inductive biases-the set of assumptions that algorithms use to make predictions for inputs they have not encountered during training. Learning on multimodal datasets presents fundamental challenges because the inductive biases can vary by data modality and graphs might not be explicitly given in the input. To address these challenges, multimodal graph AI methods combine different modalities while leveraging cross-modal dependencies using graphs. Diverse datasets are combined using graphs and fed into sophisticated multimodal architectures, specified as image-intensive, knowledge-grounded and language-intensive models. Using this categorization, we introduce a blueprint for multimodal graph learning, use it to study existing methods and provide guidelines to design new models.

Super High-Throughput Screening of Enzyme Variants by Spectral Graph Convolutional Neural Networks

Finding new enzyme variants with the desired substrate scope requires screening through a large number of potential variants. In a typical in silico enzyme engineering workflow, it is possible to scan a few thousands of variants, and gather several candidates for further screening or experimental verification. In this work, we show that a Graph Convolutional Neural Network (GCN) can be trained to predict the binding energy of combinatorial libraries of enzyme complexes using only sequence information. The GCN model uses a stack of message-passing and graph pooling layers to extract information from the protein input graph and yield a prediction. The GCN model is agnostic to the identity of the ligand, which is kept constant within the mutant libraries. Using a miniscule subset of the total combinatorial space (20⁴-20⁸ mutants) as training data, the proposed GCN model achieves a high accuracy in predicting the binding energy of unseen variants. The network's accuracy was further improved by injecting feature embeddings obtained from a language module pretrained on 10 million protein sequences. Since no structural information is needed to evaluate new variants, the deep learning algorithm is capable of scoring an enzyme variant in under 1 ms, allowing the search of billions of candidates on a single GPU.

SGPPI: structure-aware prediction of protein-protein interactions in rigorous conditions with graph convolutional network

While deep learning (DL)-based models have emerged as powerful approaches to predict protein-protein interactions (PPIs), the reliance on explicit similarity measures (e.g. sequence similarity and network neighborhood) to known interacting proteins makes these methods ineffective in dealing with novel proteins. The advent of AlphaFold2 presents a significant opportunity and also a challenge to predict PPIs in a straightforward way based on monomer structures while controlling bias from protein sequences. In this work, we established Structure and Graph-based Predictions of Protein Interactions (SGPPI), a structure-based DL framework for predicting PPIs, using the graph convolutional network. In particular, SGPPI focused on protein patches on the protein-protein binding interfaces and extracted the structural, geometric and evolutionary features from the residue contact map to predict PPIs. We demonstrated that our model outperforms traditional machine learning methods and state-of-the-art DL-based methods using non-representation-bias benchmark datasets. Moreover, our model trained on human dataset can be reliably transferred to predict yeast PPIs, indicating that SGPPI can capture converging structural features of protein interactions across various species. The implementation of SGPPI is available at https://github.com/emerson106/SGPPI.

Geometric Interaction Graph Neural Network for Predicting Protein-Ligand Binding Affinities from 3D Structures (GIGN)

Predicting protein-ligand binding affinities (PLAs) is a core problem in drug discovery. Recent advances have shown great potential in applying machine learning (ML) for PLA prediction. However, most of them omit the 3D structures of complexes and physical interactions between proteins and ligands, which are considered essential to understanding the binding mechanism. This paper proposes a geometric interaction graph neural network (GIGN) that incorporates 3D structures and physical interactions for predicting protein-ligand binding affinities. Specifically, we design a heterogeneous interaction layer that unifies covalent and noncovalent interactions into the message passing phase to learn node representations more effectively. The heterogeneous interaction layer also follows fundamental biological laws, including invariance to translations and rotations of the complexes, thus avoiding expensive data augmentation strategies. GIGN achieves state-of-the-art performance on three external test sets. Moreover, by visualizing learned representations of protein-ligand complexes, we show that the predictions of GIGN are biologically meaningful.

Explaining Black Box Drug Target Prediction Through Model Agnostic Counterfactual Samples

Many high-performance DTA deep learning models have been proposed, but they are mostly black-box and thus lack human interpretability. Explainable AI (XAI) can make DTA models more trustworthy, and allows to distill biological knowledge from the models. Counterfactual explanation is one popular approach to explaining the behaviour of a deep neural network, which works by systematically answering the question "How would the model output change if the inputs were changed in this way?". We propose a multi-agent reinforcement learning framework, Multi-Agent Counterfactual Drug-target binding Affinity (MACDA), to generate counterfactual explanations for the drug-protein complex. Our proposed framework provides human-interpretable counterfactual instances while optimizing both the input drug and target for counterfactual generation at the same time. We benchmark the proposed MACDA framework using the Davis and PDBBind dataset and find that our framework produces more parsimonious explanations with no loss in explanation validity, as measured by encoding similarity. We then present a case study involving ABL1 and Nilotinib to demonstrate how MACDA can explain the behaviour of a DTA model in the underlying substructure interaction between inputs in its prediction, revealing mechanisms that align with prior domain knowledge.

Deep Learning with Graph Convolutional Networks: An Overview and Latest Applications in Computational Intelligence

Convolutional neural networks (CNNs) have received widespread attention due to their powerful modeling capabilities and have been successfully applied in natural language processing, image recognition, and other fields. On the other hand, traditional CNN can only deal with Euclidean spatial data. In contrast, many real-life scenarios, such as transportation networks, social networks, reference networks, and so on, exist in graph data. The creation of graph convolution operators and graph pooling is at the heart of migrating CNN to graph data analysis and processing. With the advancement of the Internet and technology, graph convolution network (GCN), as an innovative technology in artificial intelligence (AI), has received more and more attention. GCN has been widely used in different fields such as image processing, intelligent recommender system, knowledge-based graph, and other areas due to their excellent characteristics in processing non-European spatial data. At the same time, communication networks have also embraced AI technology in recent years, and AI serves as the brain of the future network and realizes the comprehensive intelligence of the future grid. Many complex communication network problems can be abstracted as graph-based optimization problems and solved by GCN, thus overcoming the limitations of traditional methods. This survey briefly describes the definition of graph-based machine learning, introduces different types of graph networks, summarizes the application of GCN in various research fields, analyzes the research status, and gives the future research direction.

Predicting Potent Compounds Using a Conditional Variational Autoencoder Based upon a New Structure-Potency Fingerprint

Prediction of the potency of bioactive compounds generally relies on linear or nonlinear quantitative structure-activity relationship (QSAR) models. Nonlinear models are generated using machine learning methods. We introduce a novel approach for potency prediction that depends on a newly designed molecular fingerprint (FP) representation. This structure-potency fingerprint (SPFP) combines different modules accounting for the structural features of active compounds and their potency values in a single bit string, hence unifying structure and potency representation. This encoding enables the derivation of a conditional variational autoencoder (CVAE) using SPFPs of training compounds and apply the model to predict the SPFP potency module of test compounds using only their structure module as input. The SPFP-CVAE approach correctly predicts the potency values of compounds belonging to different activity classes with an accuracy comparable to support vector regression (SVR), representing the state-of-the-art in the field. In addition, highly potent compounds are predicted with very similar accuracy as SVR and deep neural networks.