Molecule Property Prediction Based on Spatial Graph Embedding

A deep position-encoding model for predicting olfactory perception from molecular structures and electrostatics

Predicting olfactory perceptions from odorant molecules is challenging due to the complex and potentially discontinuous nature of the perceptual space for smells. In this study, we introduce a deep learning model, Mol-PECO (Molecular Representation by Positional Encoding of Coulomb Matrix), designed to predict olfactory perceptions based on molecular structures and electrostatics. Mol-PECO learns the efficient embedding of molecules by utilizing the Coulomb matrix, which encodes atomic coordinates and charges, as an alternative of the adjacency matrix and its Laplacian eigenfunctions as positional encoding of atoms. With a comprehensive dataset of odor molecules and descriptors, Mol-PECO outperforms traditional machine learning methods using molecular fingerprints and graph neural networks based on adjacency matrices. The learned embeddings by Mol-PECO effectively capture the odor space, enabling global clustering of descriptors and local retrieval of similar odorants. This work contributes to a deeper understanding of the olfactory sense and its mechanisms.

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.

MolNet_Equi: A Chemically Intuitive, Rotation-Equivariant Graph Neural Network

Although deep-learning (DL) models suggest unprecedented prediction capabilities in tackling various chemical problems, their demonstrated tasks have so far been limited to the scalar properties including the magnitude of vectorial properties, such as molecular dipole moments. A rotation-equivariant MolNet_Equi model, proposed in this paper, understands and recognizes the molecular rotation in the 3D Euclidean space, and exhibits the ability to predict directional dipole moments in the rotation-sensitive mode, as well as showing superior performance for the prediction of scalar properties. Three consecutive operations of molecular rotationR M ${\left(R\left(M\right)\right)}$ , dipole-moment predictionφ μ R M ${\left({\phi{} }_{\mu }\left(R\left(M\right)\right)\right)}$ , and dipole-moment inverse-rotationR - 1 φ μ R M ${\left({R}^{-1}\left({\phi{} }_{\mu }\left(R\left(M\right)\right)\right)\right)}$ do not alter the original prediction of the total dipole moment of a moleculeφ μ M ${\left({\phi{} }_{\mu }\right(M\left)\right)}$ , assuring the rotational equivariance of MolNet_Equi. Furthermore, MolNet_Equi faithfully predicts the absolute direction of dipole moments given molecular poses, albeit the model has been trained only with the information on dipole-moment magnitudes, not directions. This work highlights the potential of incorporating fundamental yet crucial chemical rules and concepts into DL models, leading to the development of chemically intuitive models.

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.

Attention Mechanism-Based Graph Neural Network Model for Effective Activity Prediction of SARS-CoV-2 Main Protease Inhibitors: Application to Drug Repurposing as Potential COVID-19 Therapy

Compared to de novo drug discovery, drug repurposing provides a time-efficient way to treat coronavirus disease 19 (COVID-19) that is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 main protease (Mpro) has been proved to be an attractive drug target due to its pivotal involvement in viral replication and transcription. Here, we present a graph neural network-based deep-learning (DL) strategy to prioritize the existing drugs for their potential therapeutic effects against SARS-CoV-2 Mpro. Mpro inhibitors were represented as molecular graphs ready for graph attention network (GAT) and graph isomorphism network (GIN) modeling for predicting the inhibitory activities. The result shows that the GAT model outperforms the GIN and other competitive models and yields satisfactory predictions for unseen Mpro inhibitors, confirming its robustness and generalization. The attention mechanism of GAT enables to capture the dominant substructures and thus to realize the interpretability of the model. Finally, we applied the optimal GAT model in conjunction with molecular docking simulations to screen the Drug Repurposing Hub (DRH) database. As a result, 18 drug hits with best consensus prediction scores and binding affinity values were identified as the potential therapeutics against COVID-19. Both the extensive literature searching and evaluations on adsorption, distribution, metabolism, excretion, and toxicity (ADMET) illustrate the premium drug-likeness and pharmacokinetic properties of the drug candidates. Overall, our work not only provides an effective GAT-based DL prediction tool for inhibitory activity of SARS-CoV-2 Mpro inhibitors but also provides theoretical guidelines for drug discovery in the COVID-19 treatment.

SGNet: Sequence-Based Convolution and Ligand Graph Network for Protein Binding Affinity Prediction

Protein-ligand binding can play an important role in many fields. It is of great importance to accurately predict the binding affinity between molecules by computational methods. Most computational binding affinity methods require molecular structures. However, there are still a large number of protein molecules with known amino acid sequences whose structures have not yet been solved. To address this issue, this paper proposes a sequence-based convolution and ligand graph network, called SGNet, to fuse the molecular graph information and the amino acid sequence information. This method integrates Conjoint Triad (CT) encoding of amino acid sequence and one-dimensional convolutional neural network module to extract protein molecules, develops graph attention network to extract molecular features of ligand, and then fuses the two feature sets to predict the binding affinity between molecules from the fully connected layer. As a result, SGNet achieves good prediction performance on both KIKD and IC50 data sets, with prediction error RMSEs of 1.287 and 1.58, and correlation Pearson Rs of 0.687 and 0.592, respectively. Comparative experimental results under the same conditions showed that SGNet outperformed Kdeep and GraphDTA in predicting binding affinities between protein-ligand molecules.

Building Machine Learning Small Molecule Melting Points and Solubility Models Using CCDC Melting Points Dataset

Predicting solubility of small molecules is a very difficult undertaking due to the lack of reliable and consistent experimental solubility data. It is well known that for a molecule in a crystal lattice to be dissolved, it must, first, dissociate from the lattice and then, second, be solvated. The melting point of a compound is proportional to the lattice energy, and the octanol-water partition coefficient (log P) is a measure of the compound's solvation efficiency. The CCDC's melting point dataset of almost one hundred thousand compounds was utilized to create widely applicable machine learning models of small molecule melting points. Using the general solubility equation, the aqueous thermodynamic solubilities of the same compounds can be predicted. The global model could be easily localized by adding additional melting point measurements for a chemical series of interest.

Molecular Property Prediction by Combining LSTM and GAT

Molecular property prediction is an important direction in computer-aided drug design. In this paper, to fully explore the information from SMILE stings and graph data of molecules, we combined the SALSTM and GAT methods in order to mine the feature information of molecules from sequences and graphs. The embedding atoms are obtained through SALSTM, firstly using SMILES strings, and they are combined with graph node features and fed into the GAT to extract the global molecular representation. At the same time, data augmentation is added to enlarge the training dataset and improve the performance of the model. Finally, to enhance the interpretability of the model, the attention layers of both models are fused together to highlight the key atoms. Comparison with other graph-based and sequence-based methods, for multiple datasets, shows that our method can achieve high prediction accuracy with good generalizability.

Double-head transformer neural network for molecular property prediction

Existing molecular property prediction methods based on deep learning ignore the generalization ability of the nonlinear representation of molecular features and the reasonable assignment of weights of molecular features, making it difficult to further improve the accuracy of molecular property prediction. To solve the above problems, an end-to-end double-head transformer neural network (DHTNN) is proposed in this paper for high-precision molecular property prediction. For the data distribution characteristics of the molecular dataset, DHTNN specially designs a new activation function, beaf, which can greatly improve the generalization ability of the nonlinear representation of molecular features. A residual network is introduced in the molecular encoding part to solve the gradient explosion problem and ensure that the model can converge quickly. The transformer based on double-head attention is used to extract molecular intrinsic detail features, and the weights are reasonably assigned for predicting molecular properties with high accuracy. Our model, which was tested on the MoleculeNet [1] benchmark dataset, showed significant performance improvements over other state-of-the-art methods.

Molecular Property Prediction of Modified Gedunin Using Machine Learning

Images of molecules are often utilized in education and synthetic exploration to predict molecular characteristics. Deep learning (DL) has also had an influence on drug research, such as the interpretation of cellular images as well as the development of innovative methods for the synthesis of organic molecules. Although research in these areas has been significant, a comprehensive review of DL applications in drug development would be beyond the scope of a single Account. In this study, we will concentrate on a single major area where DL has influenced molecular design: the prediction of molecular properties of modified gedunin using machine learning (ML). AI and ML technologies are critical in drug research and development. In these other words, deep learning (DL) algorithms and artificial neural networks (ANN) have changed the field. In short, advances in AI and ML present a good potential for rational drug design and exploration, which will ultimately benefit humanity. In this paper, long short-term memory (LSTM) was used to convert a modified gedunin SMILE into a molecular image. The 2D molecular representations and their immediately visible highlights should then provide adequate data to predict the subordinate characteristics of atom design. We aim to find the properties of modified gedunin using K-means clustering; RNN-like ML tools. To support this postulation, neural network (NN) clustering based on the AI picture is used and evaluated in this study. The novel chemical developed via profound learning has long been predicted on characteristics. As a result, LSTM with RNNs allow us to predict the properties of molecules of modified gedunin. The total accuracy of the suggested model is 98.68%. The accuracy of the molecular property prediction of modified gedunin research is promising enough to evaluate extrapolation and generalization. The model suggested in this research requires just seconds or minutes to calculate, making it faster as well as more effective than existing techniques. In short, ML can be a useful tool for predicting the properties of modified gedunin molecules.

Predicting molecular properties based on the interpretable graph neural network with multistep focus mechanism

Graph neural networks based on deep learning methods have been extensively applied to the molecular property prediction because of its powerful feature learning ability and good performance. However, most of them are black boxes and cannot give the reasonable explanation about the underlying prediction mechanisms, which seriously reduce people's trust on the neural network-based prediction models. Here we proposed a novel graph neural network named iteratively focused graph network (IFGN), which can gradually identify the key atoms/groups in the molecule that are closely related to the predicted properties by the multistep focus mechanism. At the same time, the combination of the multistep focus mechanism with visualization can also generate multistep interpretations, thus allowing us to gain a deep understanding of the predictive behaviors of the model. For all studied eight datasets, the IFGN model achieved good prediction performance, indicating that the proposed multistep focus mechanism also can improve the performance of the model obviously besides increasing the interpretability of built model. For researchers to use conveniently, the corresponding website (http://graphadmet.cn/works/IFGN) was also developed and can be used free of charge.

Graph neural networks for materials science and chemistry

Machine learning plays an increasingly important role in many areas of chemistry and materials science, being used to predict materials properties, accelerate simulations, design new structures, and predict synthesis routes of new materials. Graph neural networks (GNNs) are one of the fastest growing classes of machine learning models. They are of particular relevance for chemistry and materials science, as they directly work on a graph or structural representation of molecules and materials and therefore have full access to all relevant information required to characterize materials. In this Review, we provide an overview of the basic principles of GNNs, widely used datasets, and state-of-the-art architectures, followed by a discussion of a wide range of recent applications of GNNs in chemistry and materials science, and concluding with a road-map for the further development and application of GNNs.

DENVIS: Scalable and High-Throughput Virtual Screening Using Graph Neural Networks with Atomic and Surface Protein Pocket Features

Computational methods for virtual screening can dramatically accelerate early-stage drug discovery by identifying potential hits for a specified target. Docking algorithms traditionally use physics-based simulations to address this challenge by estimating the binding orientation of a query protein-ligand pair and a corresponding binding affinity score. Over the recent years, classical and modern machine learning architectures have shown potential for outperforming traditional docking algorithms. Nevertheless, most learning-based algorithms still rely on the availability of the protein-ligand complex binding pose, typically estimated via docking simulations, which leads to a severe slowdown of the overall virtual screening process. A family of algorithms processing target information at the amino acid sequence level avoid this requirement, however, at the cost of processing protein data at a higher representation level. We introduce deep neural virtual screening (DENVIS), an end-to-end pipeline for virtual screening using graph neural networks (GNNs). By performing experiments on two benchmark databases, we show that our method performs competitively to several docking-based, machine learning-based, and hybrid docking/machine learning-based algorithms. By avoiding the intermediate docking step, DENVIS exhibits several orders of magnitude faster screening times (i.e., higher throughput) than both docking-based and hybrid models. When compared to an amino acid sequence-based machine learning model with comparable screening times, DENVIS achieves dramatically better performance. Some key elements of our approach include protein pocket modeling using a combination of atomic and surface features, the use of model ensembles, and data augmentation via artificial negative sampling during model training. In summary, DENVIS achieves competitive to state-of-the-art virtual screening performance, while offering the potential to scale to billions of molecules using minimal computational resources.

Multitype drug interaction prediction based on the deep fusion of drug features and topological relationships

Drug–drug interaction (DDI) prediction has received considerable attention from industry and academia. Most existing methods predict DDIs from drug attributes or relationships with neighbors, which does not guarantee that informative drug embeddings for prediction will be obtained. To address this limitation, we propose a multitype drug interaction prediction method based on the deep fusion of drug features and topological relationships, abbreviated DM-DDI. The proposed method adopts a deep fusion strategy to combine drug features and topologies to learn representative drug embeddings for DDI prediction. Specifically, a deep neural network model is first used on the drug feature matrix to extract feature information, while a graph convolutional network model is employed to capture structural information from the adjacency matrix. Then, we adopt delivery operations that allow the two models to exchange information between layers, as well as an attention mechanism for a weighted fusion of the two learned embeddings before the output layer. Finally, the unified drug embeddings for the downstream task are obtained. We conducted extensive experiments on real-world datasets, the experimental results demonstrated that DM-DDI achieved more accurate prediction results than state-of-the-art baselines. Furthermore, in two tasks that are more similar to real-world scenarios, DM-DDI outperformed other prediction methods for unknown drugs.

Artificial intelligence and machine-learning approaches in structure and ligand-based discovery of drugs affecting central nervous system

CNS disorders are indications with a very high unmet medical needs, relatively smaller number of available drugs, and a subpar satisfaction level among patients and caregiver. Discovery of CNS drugs is extremely expensive affair with its own unique challenges leading to extremely high attrition rates and low efficiency. With explosion of data in information age, there is hardly any aspect of life that has not been touched by data driven technologies such as artificial intelligence (AI) and machine learning (ML). Drug discovery is no exception, emergence of big data via genomic, proteomic, biological, and chemical technologies has driven pharmaceutical giants to collaborate with AI oriented companies to revolutionise drug discovery, with the goal of increasing the efficiency of the process. In recent years many examples of innovative applications of AI and ML techniques in CNS drug discovery has been reported. Research on therapeutics for diseases such as schizophrenia, Alzheimer's and Parkinsonism has been provided with a new direction and thrust from these developments. AI and ML has been applied to both ligand-based and structure-based drug discovery and design of CNS therapeutics. In this review, we have summarised the general aspects of AI and ML from the perspective of drug discovery followed by a comprehensive coverage of the recent developments in the applications of AI/ML techniques in CNS drug discovery.

Multi-channel GCN ensembled machine learning model for molecular aqueous solubility prediction on a clean dataset

This study constructed a new aqueous solubility dataset and a solubility regression model which was ensembled by GCN and machine learning models. Aqueous solubility is a key physiochemical property of small molecules in drug discovery. In the past few decades, there have been many studies about solubility prediction. However, many of these studies have high root mean squared error (RMSE). Meanwhile, their dataset always contains salt compounds and solubility data obtained from different experimental conditions. In this paper, we constructed a clean dataset with 2609 compounds, which was small but contains only solubility records without salts at the same temperatures (25 °C). Here, we applied graph convolutional neural network (GCN) to construct an aqueous solubility prediction model. To enhance the performance of the model, the molecular MACCS key fingerprints and physiochemical descriptors were also combined with the GCN model to build a multi-channel model. Additionally, the authors also built two machine learning models (support vector regression and gradient boost decision tree) and assembled them to the GCN model to improve the root mean squared error (RMSE = 0.665). Finally, comparative experiments have shown that our framework achieved the best performance on ESOL dataset (RMSE_val = 0.56, RMSE_test = 0.44) and surpassed four established software on aqueous solubility prediction of new compounds.

Multiphysical graph neural network (MP-GNN) for COVID-19 drug design

Graph neural networks (GNNs) are the most promising deep learning models that can revolutionize non-Euclidean data analysis. However, their full potential is severely curtailed by poorly represented molecular graphs and features. Here, we propose a multiphysical graph neural network (MP-GNN) model based on the developed multiphysical molecular graph representation and featurization. All kinds of molecular interactions, between different atom types and at different scales, are systematically represented by a series of scale-specific and element-specific graphs with distance-related node features. From these graphs, graph convolution network (GCN) models are constructed with specially designed weight-sharing architectures. Base learners are constructed from GCN models from different elements at different scales, and further consolidated together using both one-scale and multi-scale ensemble learning schemes. Our MP-GNN has two distinct properties. First, our MP-GNN incorporates multiscale interactions using more than one molecular graph. Atomic interactions from various different scales are not modeled by one specific graph (as in traditional GNNs), instead they are represented by a series of graphs at different scales. Second, it is free from the complicated feature generation process as in conventional GNN methods. In our MP-GNN, various atom interactions are embedded into element-specific graph representations with only distance-related node features. A unique GNN architecture is designed to incorporate all the information into a consolidated model. Our MP-GNN has been extensively validated on the widely used benchmark test datasets from PDBbind, including PDBbind-v2007, PDBbind-v2013 and PDBbind-v2016. Our model can outperform all existing models as far as we know. Further, our MP-GNN is used in coronavirus disease 2019 drug design. Based on a dataset with 185 complexes of inhibitors for severe acute respiratory syndrome coronavirus (SARS-CoV/SARS-CoV-2), we evaluate their binding affinities using our MP-GNN. It has been found that our MP-GNN is of high accuracy. This demonstrates the great potential of our MP-GNN for the screening of potential drugs for SARS-CoV-2. Availability: The Multiphysical graph neural network (MP-GNN) model can be found in https://github.com/Alibaba-DAMO-DrugAI/MGNN. Additional data or code will be available upon reasonable request.

AI-Aided Design of Novel Targeted Covalent Inhibitors against SARS-CoV-2

The drug repurposing of known approved drugs (e.g., lopinavir/ritonavir) has failed to treat SARS-CoV-2-infected patients. Therefore, it is important to generate new chemical entities against this virus. As a critical enzyme in the lifecycle of the coronavirus, the 3C-like main protease (3CLpro or Mpro) is the most attractive target for antiviral drug design. Based on a recently solved structure (PDB ID: 6LU7), we developed a novel advanced deep Q-learning network with a fragment-based drug design (ADQN-FBDD) for generating potential lead compounds targeting SARS-CoV-2 3CLpro. We obtained a series of derivatives from the lead compounds based on our structure-based optimization policy (SBOP). All of the 47 lead compounds obtained directly with our AI model and related derivatives based on the SBOP are accessible in our molecular library. These compounds can be used as potential candidates by researchers to develop drugs against SARS-CoV-2.

AI-Aided Design of Novel Targeted Covalent Inhibitors against SARS-CoV-2

The drug repurposing of known approved drugs (e.g., lopinavir/ritonavir) has failed to treat SARS-CoV-2-infected patients. Therefore, it is important to generate new chemical entities against this virus. As a critical enzyme in the lifecycle of the coronavirus, the 3C-like main protease (3CLpro or Mpro) is the most attractive target for antiviral drug design. Based on a recently solved structure (PDB ID: 6LU7), we developed a novel advanced deep Q-learning network with a fragment-based drug design (ADQN-FBDD) for generating potential lead compounds targeting SARS-CoV-2 3CLpro. We obtained a series of derivatives from the lead compounds based on our structure-based optimization policy (SBOP). All of the 47 lead compounds obtained directly with our AI model and related derivatives based on the SBOP are accessible in our molecular library. These compounds can be used as potential candidates by researchers to develop drugs against SARS-CoV-2.

Machine Learning of Reaction Properties via Learned Representations of the Condensed Graph of Reaction

The estimation of chemical reaction properties such as activation energies, rates, or yields is a central topic of computational chemistry. In contrast to molecular properties, where machine learning approaches such as graph convolutional neural networks (GCNNs) have excelled for a wide variety of tasks, no general and transferable adaptations of GCNNs for reactions have been developed yet. We therefore combined a popular cheminformatics reaction representation, the so-called condensed graph of reaction (CGR), with a recent GCNN architecture to arrive at a versatile, robust, and compact deep learning model. The CGR is a superposition of the reactant and product graphs of a chemical reaction and thus an ideal input for graph-based machine learning approaches. The model learns to create a data-driven, task-dependent reaction embedding that does not rely on expert knowledge, similar to current molecular GCNNs. Our approach outperforms current state-of-the-art models in accuracy, is applicable even to imbalanced reactions, and possesses excellent predictive capabilities for diverse target properties, such as activation energies, reaction enthalpies, rate constants, yields, or reaction classes. We furthermore curated a large set of atom-mapped reactions along with their target properties, which can serve as benchmark data sets for future work. All data sets and the developed reaction GCNN model are available online, free of charge, and open source.

Assisting Multitargeted Ligand Affinity Prediction of Receptor Tyrosine Kinases Associated Nonsmall Cell Lung Cancer Treatment with Multitasking Principal Neighborhood Aggregation

A multitargeted therapeutic approach with hybrid drugs is a promising strategy to enhance anticancer efficiency and overcome drug resistance in nonsmall cell lung cancer (NSCLC) treatment. Estimating affinities of small molecules against targets of interest typically proceeds as a preliminary action for recent drug discovery in the pharmaceutical industry. In this investigation, we employed machine learning models to provide a computationally affordable means for computer-aided screening to accelerate the discovery of potential drug compounds. In particular, we introduced a quantitative structure–activity-relationship (QSAR)-based multitask learning model to facilitate an in silico screening system of multitargeted drug development. Our method combines a recently developed graph-based neural network architecture, principal neighborhood aggregation (PNA), with a descriptor-based deep neural network supporting synergistic utilization of molecular graph and fingerprint features. The model was generated by more than ten-thousands affinity-reported ligands of seven crucial receptor tyrosine kinases in NSCLC from two public data sources. As a result, our multitask model demonstrated better performance than all other benchmark models, as well as achieving satisfying predictive ability regarding applicable QSAR criteria for most tasks within the model’s applicability. Since our model could potentially be a screening tool for practical use, we have provided a model implementation platform with a tutorial that is freely accessible hence, advising the first move in a long journey of cancer drug development.

Flexible Dual-Branched Message-Passing Neural Network for a Molecular Property Prediction

A molecule is a complex of heterogeneous components, and the spatial arrangements of these components determine the whole molecular properties and characteristics. With the advent of deep learning in computational chemistry, several studies have focused on how to predict molecular properties based on molecular configurations. MA message-passing neural network provides an effective framework for capturing molecular geometric features with the perspective of a molecule as a graph. However, most of these studies assumed that all heterogeneous molecular features, such as atomic charge, bond length, or other geometric features, always contribute equivalently to the target prediction, regardless of the task type. In this study, we propose a dual-branched neural network for molecular property prediction based on both the message-passing framework and standard multilayer perceptron neural networks. Our model learns heterogeneous molecular features with different scales, which are trained flexibly according to each prediction target. In addition, we introduce a discrete branch to learn single-atom features without local aggregation, apart from message-passing steps. We verify that this novel structure can improve the model performance. The proposed model outperforms other recent models with sparser representations. Our experimental results indicate that, in the chemical property prediction tasks, the diverse chemical nature of targets should be carefully considered for both model performance and generalizability. Finally, we provide the intuitive analysis between the experimental results and the chemical meaning of the target.

Authentic Intelligent Machine for Scaling Driven Discovery: A Case for Chiral Quantum Dots

The scaling laws have long been used as evidence of science where many fundamental physics laws emerge. As emerging nanomaterials, quantum dots are also sensitive to scaling because of their strong size effect. In this work, we developed the chiral dielectric theory based on the exciton absorption mechanism to explain the increment of the dielectric constant from chirality via its dimensionality. To help researchers discover and develop scaling relevant theories, the Authentic Intelligent Machine (AIM) protocol was developed to generate and interpret experimental data in an analytical and scaling-oriented manner. We show how the AIM protocol interprets spectra such as transient absorption data of chiral quantum dots with theories, where discrepancies concerning the dielectric constant were discovered. Examples for applying the AIM protocol on other spectra, such as absorption spectra and photoluminescence spectra, are also given.

Graph Neural Networks as a Potential Tool in Improving Virtual Screening Programs

Despite the increasing number of pharmaceutical companies, university laboratories and funding, less than one percent of initially researched drugs enter the commercial market. In this context, virtual screening (VS) has gained much attention due to several advantages, including timesaving, reduced reagent and consumable costs and the performance of selective analyses regarding the affinity between test molecules and pharmacological targets. Currently, VS is based mainly on algorithms that apply physical and chemistry principles and quantum mechanics to estimate molecule affinities and conformations, among others. Nevertheless, VS has not reached the expected results concerning the improvement of market-approved drugs, comprising less than twenty drugs that have reached this goal to date. In this context, graph neural networks (GNN), a recent deep-learning subtype, may comprise a powerful tool to improve VS results concerning natural products that may be used both simultaneously with standard algorithms or isolated. This review discusses the pros and cons of GNN applied to VS and the future perspectives of this learnable algorithm, which may revolutionize drug discovery if certain obstacles concerning spatial coordinates and adequate datasets, among others, can be overcome.

Molecular persistent spectral image (Mol-PSI) representation for machine learning models in drug design

Artificial intelligence (AI)-based drug design has great promise to fundamentally change the landscape of the pharmaceutical industry. Even though there are great progress from handcrafted feature-based machine learning models, 3D convolutional neural networks (CNNs) and graph neural networks, effective and efficient representations that characterize the structural, physical, chemical and biological properties of molecular structures and interactions remain to be a great challenge. Here, we propose an equal-sized molecular 2D image representation, known as the molecular persistent spectral image (Mol-PSI), and combine it with CNN model for AI-based drug design. Mol-PSI provides a unique one-to-one image representation for molecular structures and interactions. In general, deep models are empowered to achieve better performance with systematically organized representations in image format. A well-designed parallel CNN architecture for adapting Mol-PSIs is developed for protein-ligand binding affinity prediction. Our results, for the three most commonly used databases, including PDBbind-v2007, PDBbind-v2013 and PDBbind-v2016, are better than all traditional machine learning models, as far as we know. Our Mol-PSI model provides a powerful molecular representation that can be widely used in AI-based drug design and molecular data analysis.

A Comparative Study of Marginalized Graph Kernel and Message-Passing Neural Network

This work proposes a state-of-the-art hybrid kernel to calculate molecular similarity. Combined with Gaussian process models, the performance of the hybrid kernel in predicting molecular properties is comparable to that of the directed message-passing neural network (D-MPNN). The hybrid kernel consists of a marginalized graph kernel (MGK) and a radial basis function (RBF) kernel that operate on molecular graphs and global molecular features, respectively. Bayesian optimization was used to obtain the optimal hyperparameters for both models. The comparisons are performed on 11 publicly available data sets. Our results show that their performances are similar, their prediction errors are correlated, and the ensemble predictions of the two models perform better than either of them. Through principal component analysis, we found that the molecular embeddings of the hybrid kernel and the D-MPNN are also similar. The advantage of D-MPNN lies in the computational efficiency and scalability of large-scale data, while the advantage of the graph kernel models lies in the accurate uncertainty quantification.

Benchmarking Machine Learning Models for Polymer Informatics: An Example of Glass Transition Temperature

In the field of polymer informatics, utilizing machine learning (ML) techniques to evaluate the glass transition temperature T_g and other properties of polymers has attracted extensive attention. This data-centric approach is much more efficient and practical than the laborious experimental measurements when encountered a daunting number of polymer structures. Various ML models are demonstrated to perform well for T_g prediction. Nevertheless, they are trained on different data sets, using different structure representations, and based on different feature engineering methods. Thus, the critical question arises on selecting a proper ML model to better handle the T_g prediction with generalization ability. To provide a fair comparison of different ML techniques and examine the key factors that affect the model performance, we carry out a systematic benchmark study by compiling 79 different ML models and training them on a large and diverse data set. The three major components in setting up an ML model are structure representations, feature representations, and ML algorithms. In terms of polymer structure representation, we consider the polymer monomer, repeat unit, and oligomer with longer chain structure. Based on that feature, representation is calculated, including Morgan fingerprinting with or without substructure frequency, RDKit descriptors, molecular embedding, molecular graph, etc. Afterward, the obtained feature input is trained using different ML algorithms, such as deep neural networks, convolutional neural networks, random forest, support vector machine, LASSO regression, and Gaussian process regression. We evaluate the performance of these ML models using a holdout test set and an extra unlabeled data set from high-throughput molecular dynamics simulation. The ML model's generalization ability on an unlabeled data set is especially focused, and the model's sensitivity to topology and the molecular weight of polymers is also taken into consideration. This benchmark study provides not only a guideline for the T_g prediction task but also a useful reference for other polymer informatics tasks.

Artificial intelligence in drug discovery: applications and techniques

Artificial intelligence (AI) has been transforming the practice of drug discovery in the past decade. Various AI techniques have been used in many drug discovery applications, such as virtual screening and drug design. In this survey, we first give an overview on drug discovery and discuss related applications, which can be reduced to two major tasks, i.e. molecular property prediction and molecule generation. We then present common data resources, molecule representations and benchmark platforms. As a major part of the survey, AI techniques are dissected into model architectures and learning paradigms. To reflect the technical development of AI in drug discovery over the years, the surveyed works are organized chronologically. We expect that this survey provides a comprehensive review on AI in drug discovery. We also provide a GitHub repository with a collection of papers (and codes, if applicable) as a learning resource, which is regularly updated.

Graph representation learning in bioinformatics: trends, methods and applications

Graph is a natural data structure for describing complex systems, which contains a set of objects and relationships. Ubiquitous real-life biomedical problems can be modeled as graph analytics tasks. Machine learning, especially deep learning, succeeds in vast bioinformatics scenarios with data represented in Euclidean domain. However, rich relational information between biological elements is retained in the non-Euclidean biomedical graphs, which is not learning friendly to classic machine learning methods. Graph representation learning aims to embed graph into a low-dimensional space while preserving graph topology and node properties. It bridges biomedical graphs and modern machine learning methods and has recently raised widespread interest in both machine learning and bioinformatics communities. In this work, we summarize the advances of graph representation learning and its representative applications in bioinformatics. To provide a comprehensive and structured analysis and perspective, we first categorize and analyze both graph embedding methods (homogeneous graph embedding, heterogeneous graph embedding, attribute graph embedding) and graph neural networks. Furthermore, we summarize their representative applications from molecular level to genomics, pharmaceutical and healthcare systems level. Moreover, we provide open resource platforms and libraries for implementing these graph representation learning methods and discuss the challenges and opportunities of graph representation learning in bioinformatics. This work provides a comprehensive survey of emerging graph representation learning algorithms and their applications in bioinformatics. It is anticipated that it could bring valuable insights for researchers to contribute their knowledge to graph representation learning and future-oriented bioinformatics studies.

Prediction of Molecular Properties Using Molecular Topographic Map

Prediction of molecular properties plays a critical role towards rational drug design. In this study, the Molecular Topographic Map (MTM) is proposed, which is a two-dimensional (2D) map that can be used to represent a molecule. An MTM is generated from the atomic features set of a molecule using generative topographic mapping and is then used as input data for analyzing structure-property/activity relationships. In the visualization and classification of 20 amino acids, differences of the amino acids can be visually confirmed from and revealed by hierarchical clustering with a similarity matrix of their MTMs. The prediction of molecular properties was performed on the basis of convolutional neural networks using MTMs as input data. The performance of the predictive models using MTM was found to be equal to or better than that using Morgan fingerprint or MACCS keys. Furthermore, data augmentation of MTMs using mixup has improved the prediction performance. Since molecules converted to MTMs can be treated like 2D images, they can be easily used with existing neural networks for image recognition and related technologies. MTM can be effectively utilized to predict molecular properties of small molecules to aid drug discovery research.

XGraphBoost: Extracting Graph Neural Network-Based Features for a Better Prediction of Molecular Properties

Determining the properties of chemical molecules is essential for screening candidates similar to a specific drug. These candidate molecules are further evaluated for their target binding affinities, side effects, target missing probabilities, etc. Conventional machine learning algorithms demonstrated satisfying prediction accuracies of molecular properties. A molecule cannot be directly loaded into a machine learning model, and a set of engineered features needs to be designed and calculated from a molecule. Such hand-crafted features rely heavily on the experiences of the investigating researchers. The concept of graph neural networks (GNNs) was recently introduced to describe the chemical molecules. The features may be automatically and objectively extracted from the molecules through various types of GNNs, e.g., GCN (graph convolution network), GGNN (gated graph neural network), DMPNN (directed message passing neural network), etc. However, the training of a stable GNN model requires a huge number of training samples and a large amount of computing power, compared with the conventional machine learning strategies. This study proposed the integrated framework XGraphBoost to extract the features using a GNN and build an accurate prediction model of molecular properties using the classifier XGBoost. The proposed framework XGraphBoost fully inherits the merits of the GNN-based automatic molecular feature extraction and XGBoost-based accurate prediction performance. Both classification and regression problems were evaluated using the framework XGraphBoost. The experimental results strongly suggest that XGraphBoost may facilitate the efficient and accurate predictions of various molecular properties. The source code is freely available to academic users at https://github.com/chenxiaowei-vincent/XGraphBoost.git.

Deep Learning Optical Spectroscopy Based on Experimental Database: Potential Applications to Molecular Design

Accurate and reliable prediction of the optical and photophysical properties of organic compounds is important in various research fields. Here, we developed deep learning (DL) optical spectroscopy using a DL model and experimental database to predict seven optical and photophysical properties of organic compounds, namely, the absorption peak position and bandwidth, extinction coefficient, emission peak position and bandwidth, photoluminescence quantum yield (PLQY), and emission lifetime. Our DL model included the chromophore-solvent interaction to account for the effect of local environments on the optical and photophysical properties of organic compounds and was trained using an experimental database of 30 094 chromophore/solvent combinations. Our DL optical spectroscopy made it possible to reliably and quickly predict the aforementioned properties of organic compounds in solution, gas phase, film, and powder with the root mean squared errors of 26.6 and 28.0 nm for absorption and emission peak positions, 603 and 532 cm-1 for absorption and emission bandwidths, and 0.209, 0.371, and 0.262 for the logarithm of the extinction coefficient, PLQY, and emission lifetime, respectively. Finally, we demonstrated how a blue emitter with desired optical and photophysical properties could be efficiently virtually screened and developed by DL optical spectroscopy. DL optical spectroscopy can be efficiently used for developing chromophores and fluorophores in various research areas.

Persistent spectral hypergraph based machine learning (PSH-ML) for protein-ligand binding affinity prediction

Molecular descriptors are essential to not only quantitative structure activity/property relationship (QSAR/QSPR) models, but also machine learning based chemical and biological data analysis. In this paper, we propose persistent spectral hypergraph (PSH) based molecular descriptors or fingerprints for the first time. Our PSH-based molecular descriptors are used in the characterization of molecular structures and interactions, and further combined with machine learning models, in particular gradient boosting tree (GBT), for protein-ligand binding affinity prediction. Different from traditional molecular descriptors, which are usually based on molecular graph models, a hypergraph-based topological representation is proposed for protein-ligand interaction characterization. Moreover, a filtration process is introduced to generate a series of nested hypergraphs in different scales. For each of these hypergraphs, its eigen spectrum information can be obtained from the corresponding (Hodge) Laplacain matrix. PSH studies the persistence and variation of the eigen spectrum of the nested hypergraphs during the filtration process. Molecular descriptors or fingerprints can be generated from persistent attributes, which are statistical or combinatorial functions of PSH, and combined with machine learning models, in particular, GBT. We test our PSH-GBT model on three most commonly used datasets, including PDBbind-2007, PDBbind-2013 and PDBbind-2016. Our results, for all these databases, are better than all existing machine learning models with traditional molecular descriptors, as far as we know.

A spatial-temporal gated attention module for molecular property prediction based on molecular geometry

MOTIVATION

Geometry-based properties and characteristics of drug molecules play an important role in drug development for virtual screening in computational chemistry. The 3D characteristics of molecules largely determine the properties of the drug and the binding characteristics of the target. However, most of the previous studies focused on 1D or 2D molecular descriptors while ignoring the 3D topological structure, thereby degrading the performance of molecule-related prediction. Because it is very time-consuming to use dynamics to simulate molecular 3D conformer, we aim to use machine learning to represent 3D molecules by using the generated 3D molecular coordinates from the 2D structure.

RESULTS

We proposed Drug3D-Net, a novel deep neural network architecture based on the spatial geometric structure of molecules for predicting molecular properties. It is grid-based 3D convolutional neural network with spatial-temporal gated attention module, which can extract the geometric features for molecular prediction tasks in the process of convolution. The effectiveness of Drug3D-Net is verified on the public molecular datasets. Compared with other deep learning methods, Drug3D-Net shows superior performance in predicting molecular properties and biochemical activities.

AVAILABILITY AND IMPLEMENTATION

https://github.com/anny0316/Drug3D-Net.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/bib.

A Novel Graph Neural Network Methodology to Investigate Dihydroorotate Dehydrogenase Inhibitors in Small Cell Lung Cancer

Small cell lung cancer (SCLC) is a particularly aggressive tumor subtype, and dihydroorotate dehydrogenase (DHODH) has been demonstrated to be a therapeutic target for SCLC. Network pharmacology analysis and virtual screening were utilized to find out related proteins and investigate candidates with high docking capacity to multiple targets. Graph neural networks (GNNs) and machine learning were used to build reliable predicted models. We proposed a novel concept of multi-GNNs, and then built three multi-GNN models called GIAN, GIAT, and SGCA, which achieved satisfactory results in our dataset containing 532 molecules with all R^{^2} values greater than 0.92 on the training set and higher than 0.8 on the test set. Compared with machine learning algorithms, random forest (RF), and support vector regression (SVR), multi-GNNs had a better modeling effect and higher precision. Furthermore, the long-time 300 ns molecular dynamics simulation verified the stability of the protein–ligand complexes. The result showed that ZINC8577218, ZINC95618747, and ZINC4261765 might be the potentially potent inhibitors for DHODH. Multi-GNNs show great performance in practice, making them a promising field for future research. We therefore suggest that this novel concept of multi-GNNs is a promising protocol for drug discovery.

Large-Scale Modeling of Multispecies Acute Toxicity End Points Using Consensus of Multitask Deep Learning Methods

Computational methods to predict molecular properties regarding safety and toxicology represent alternative approaches to expedite drug development, screen environmental chemicals, and thus significantly reduce associated time and costs. There is a strong need and interest in the development of computational methods that yield reliable predictions of toxicity, and many approaches, including the recently introduced deep neural networks, have been leveraged towards this goal. Herein, we report on the collection, curation, and integration of data from the public data sets that were the source of the ChemIDplus database for systemic acute toxicity. These efforts generated the largest publicly available such data set comprising > 80,000 compounds measured against a total of 59 acute systemic toxicity end points. This data was used for developing multiple single- and multitask models utilizing random forest, deep neural networks, convolutional, and graph convolutional neural network approaches. For the first time, we also reported the consensus models based on different multitask approaches. To the best of our knowledge, prediction models for 36 of the 59 end points have never been published before. Furthermore, our results demonstrated a significantly better performance of the consensus model obtained from three multitask learning approaches that particularly predicted the 29 smaller tasks (less than 300 compounds) better than other models developed in the study. The curated data set and the developed models have been made publicly available at https://github.com/ncats/ld50-multitask, https://predictor.ncats.io/, and https://cactus.nci.nih.gov/download/acute-toxicity-db (data set only) to support regulatory and research applications.

An end-to-end heterogeneous graph representation learning-based framework for drug-target interaction prediction

Accurately identifying potential drug-target interactions (DTIs) is a key step in drug discovery. Although many related experimental studies have been carried out for identifying DTIs in the past few decades, the biological experiment-based DTI identification is still timeconsuming and expensive. Therefore, it is of great significance to develop effective computational methods for identifying DTIs. In this paper, we develop a novel 'end-to-end' learning-based framework based on heterogeneous 'graph' convolutional networks for 'DTI' prediction called end-to-end graph (EEG)-DTI. Given a heterogeneous network containing multiple types of biological entities (i.e. drug, protein, disease, side-effect), EEG-DTI learns the low-dimensional feature representation of drugs and targets using a graph convolutional networks-based model and predicts DTIs based on the learned features. During the training process, EEG-DTI learns the feature representation of nodes in an end-to-end mode. The evaluation test shows that EEG-DTI performs better than existing state-of-art methods. The data and source code are available at: https://github.com/MedicineBiology-AI/EEG-DTI.

Prediction of pharmacological activities from chemical structures with graph convolutional neural networks

Many therapeutic drugs are compounds that can be represented by simple chemical structures, which contain important determinants of affinity at the site of action. Recently, graph convolutional neural network (GCN) models have exhibited excellent results in classifying the activity of such compounds. For models that make quantitative predictions of activity, more complex information has been utilized, such as the three-dimensional structures of compounds and the amino acid sequences of their respective target proteins. As another approach, we hypothesized that if sufficient experimental data were available and there were enough nodes in hidden layers, a simple compound representation would quantitatively predict activity with satisfactory accuracy. In this study, we report that GCN models constructed solely from the two-dimensional structural information of compounds demonstrated a high degree of activity predictability against 127 diverse targets from the ChEMBL database. Using the information entropy as a metric, we also show that the structural diversity had less effect on the prediction performance. Finally, we report that virtual screening using the constructed model identified a new serotonin transporter inhibitor with activity comparable to that of a marketed drug in vitro and exhibited antidepressant effects in behavioural studies.

The application potential of machine learning and genomics for understanding natural product diversity, chemistry, and therapeutic translatability

Covering: up to the end of 2020. The machine learning field can be defined as the study and application of algorithms that perform classification and prediction tasks through pattern recognition instead of explicitly defined rules. Among other areas, machine learning has excelled in natural language processing. As such methods have excelled at understanding written languages (e.g. English), they are also being applied to biological problems to better understand the "genomic language". In this review we focus on recent advances in applying machine learning to natural products and genomics, and how those advances are improving our understanding of natural product biology, chemistry, and drug discovery. We discuss machine learning applications in genome mining (identifying biosynthetic signatures in genomic data), predictions of what structures will be created from those genomic signatures, and the types of activity we might expect from those molecules. We further explore the application of these approaches to data derived from complex microbiomes, with a focus on the human microbiome. We also review challenges in leveraging machine learning approaches in the field, and how the availability of other "omics" data layers provides value. Finally, we provide insights into the challenges associated with interpreting machine learning models and the underlying biology and promises of applying machine learning to natural product drug discovery. We believe that the application of machine learning methods to natural product research is poised to accelerate the identification of new molecular entities that may be used to treat a variety of disease indications.

A compact review of molecular property prediction with graph neural networks

As graph neural networks are becoming more and more powerful and useful in the field of drug discovery, many pharmaceutical companies are getting interested in utilizing these methods for their own in-house frameworks. This is especially compelling for tasks such as the prediction of molecular properties which is often one of the most crucial tasks in computer-aided drug discovery workflows. The immense hype surrounding these kinds of algorithms has led to the development of many different types of promising architectures and in this review we try to structure this highly dynamic field of AI-research by collecting and classifying 80 GNNs that have been used to predict more than 20 molecular properties using 48 different datasets.

Accurate predictions of aqueous solubility of drug molecules via the multilevel graph convolutional network (MGCN) and SchNet architectures

Deep learning based methods have been widely applied to predict various kinds of molecular properties in the pharmaceutical industry with increasingly more success. In this study, we propose two novel models for aqueous solubility predictions, based on the Multilevel Graph Convolutional Network (MGCN) and SchNet architectures, respectively. The advantage of the MGCN lies in the fact that it could extract the graph features of the target molecules directly from the (3D) structural information; therefore, it doesn't need to rely on a lot of intra-molecular descriptors to learn the features, which are of significance for accurate predictions of the molecular properties. The SchNet performs well in modelling the interatomic interactions inside a molecule, and such a deep learning architecture is also capable of extracting structural information and further predicting the related properties. The actual accuracy of these two novel approaches was systematically benchmarked with four different independent datasets. We found that both the MGCN and SchNet models performed well for aqueous solubility predictions. In the future, we believe such promising predictive models will be applicable to enhancing the efficiency of the screening, crystallization and delivery of drug molecules, essentially as a useful tool to promote the development of molecular pharmaceutics.

Inductive transfer learning for molecular activity prediction: Next-Gen QSAR Models with MolPMoFiT

Deep neural networks can directly learn from chemical structures without extensive, user-driven selection of descriptors in order to predict molecular properties/activities with high reliability. But these approaches typically require large training sets to learn the endpoint-specific structural features and ensure reasonable prediction accuracy. Even though large datasets are becoming the new normal in drug discovery, especially when it comes to high-throughput screening or metabolomics datasets, one should also consider smaller datasets with challenging endpoints to model and forecast. Thus, it would be highly relevant to better utilize the tremendous compendium of unlabeled compounds from publicly-available datasets for improving the model performances for the user’s particular series of compounds. In this study, we propose the Molecular Prediction Model Fine-Tuning (MolPMoFiT) approach, an effective transfer learning method based on self-supervised pre-training + task-specific fine-tuning for QSPR/QSAR modeling. A large-scale molecular structure prediction model is pre-trained using one million unlabeled molecules from ChEMBL in a self-supervised learning manner, and can then be fine-tuned on various QSPR/QSAR tasks for smaller chemical datasets with specific endpoints. Herein, the method is evaluated on four benchmark datasets (lipophilicity, FreeSolv, HIV, and blood–brain barrier penetration). The results showed the method can achieve strong performances for all four datasets compared to other state-of-the-art machine learning modeling techniques reported in the literature so far.

AI-aided design of novel targeted covalent inhibitors against SARS-CoV-2

The focused drug repurposing of known approved drugs (such as lopinavir/ritonavir) has been reported failed for curing SARS-CoV-2 infected patients. It is urgent to generate new chemical entities against this virus. As a key enzyme in the life-cycle of coronavirus, the 3C-like main protease (3CL ^pro or M ^pro ) is the most attractive for antiviral drug design. Based on a recently solved structure (PDB ID: 6LU7), we developed a novel advanced deep Q-learning network with the fragment-based drug design (ADQN-FBDD) for generating potential lead compounds targeting SARS-CoV-2 3CL ^pro . We obtained a series of derivatives from those lead compounds by our structure-based optimization policy (SBOP). All the 47 lead compounds directly from our AI-model and related derivatives based on SBOP are accessible in our molecular library at https://github.com/tbwxmu/2019-nCov . These compounds can be used as potential candidates for researchers in their development of drugs against SARS-CoV-2.

A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility

Efficient and accurate prediction of molecular properties, such as lipophilicity and solubility, is highly desirable for rational compound design in chemical and pharmaceutical industries. To this end, we build and apply a graph-neural-network framework called self-attention-based message-passing neural network (SAMPN) to study the relationship between chemical properties and structures in an interpretable way. The main advantages of SAMPN are that it directly uses chemical graphs and breaks the black-box mold of many machine/deep learning methods. Specifically, its attention mechanism indicates the degree to which each atom of the molecule contributes to the property of interest, and these results are easily visualized. Further, SAMPN outperforms random forests and the deep learning framework MPN from Deepchem. In addition, another formulation of SAMPN (Multi-SAMPN) can simultaneously predict multiple chemical properties with higher accuracy and efficiency than other models that predict one specific chemical property. Moreover, SAMPN can generate chemically visible and interpretable results, which can help researchers discover new pharmaceuticals and materials. The source code of the SAMPN prediction pipeline is freely available at Github (https://github.com/tbwxmu/SAMPN).