A Deep Learning Drug-Target Binding Affinity Prediction Based on Compound Microstructure and Its Application in COVID-19 Drug Screening

Drug target relationship(DTR)prediction is a rapidly evolving area of research in com-putational drug discovery.Despite recent advances in computational solutions that have overcome the challenges of in vitro and in vivo experiments,most computational methods still focus on binary classification.They ignore the importance of binding affinity,which correctly distinguishes between on-targets and off-targets.In this study,we propose a deep learning model based on the microstruc-ture of compounds and proteins to predict drug-target binding affinity(DTA),which utilizes topo-logical structure information of drug molecules and sequence semantic information of proteins.In this model,graph attention network(GAT)is used to capture the deep features of the compound molecular graph,and bidirectional long short-term memory(BiLSTM)network is used to extract the protein sequence features,and the pharmacological context of DTA is obtained by combining the two.The results show that the proposed model has achieved superior performance in both cor-rectly predicting the value of interaction strength and correctly discriminating the ranking of bind-ing strength compared to the state-of-the-art baselines.A case study experiment on COVID-19 con-firms that the proposed DTA model can be used as an effective pre-screening tool in drug discovery.

DeepDrug:A general graph-based deep learning framework for drug-drug interactions and drug-target interactions prediction

Background:Computational approaches for accurate prediction of drug interactions,such as drug-drug interactions(DDIs)and drug-target interactions(DTIs),are highly demanded for biochemical researchers.Despite the fact that many methods have been proposed and developed to predict DDIs and DTIs respectively,their success is still limited due to a lack of systematic evaluation of the intrinsic properties embedded in the corresponding chemical structure.Methods:In this paper,we develop DeepDrug,a deep learning framework for overcoming the above limitation by using residual graph convolutional networks(Res-GCNs)and convolutional networks(CNNs)to learn the comprehensive structure-and sequence-based representations of drugs and proteins.Results:DeepDrug outperforms state-of-the-art methods in a series of systematic experiments,including binary-class DDIs,multi-class/multi-label DDIs,binary-class DTIs classification and DTIs regression tasks.Furthermore,we visualize the structural features learned by DeepDrug Res-GCN module,which displays compatible and accordant patterns in chemical properties and drug categories,providing additional evidence to support the strong predictive power of DeepDrug.Ultimately,we apply DeepDrug to perform drug repositioning on the whole DrugBank database to discover the potential drug candidates against SARS-CoV-2,where 7 out of 10 top-ranked drugs are reported to be repurposed to potentially treat coronavirus disease 2019(COVID-19).Conclusions:To sum up,we believe that DeepDrug is an efficient tool in accurate prediction of DDIs and DTIs and provides a promising insight in understanding the underlying mechanism of these biochemical relations.

Relational Topology-based Heterogeneous Network Embedding for Predicting Drug-Target Interactions

Predicting interactions between drugs and target proteins has become an essential task in the drug discovery process.Although the method of validation via wet-lab experiments has become available,experimental methods for drug-target interaction(DTI)identification remain either time consuming or heavily dependent on domain expertise.Therefore,various computational models have been proposed to predict possible interactions between drugs and target proteins.However,most prediction methods do not consider the topological structures characteristics of the relationship.In this paper,we propose a relational topologybased heterogeneous network embedding method to predict drug-target interactions,abbreviated as RTHNE_DTI.We first construct a heterogeneous information network based on the interaction between different types of nodes,to enhance the ability of association discovery by fully considering the topology of the network.Then drug and target protein nodes can be represented by the other types of nodes.According to the different topological structure of the relationship between the nodes,we divide the relationship in the heterogeneous network into two categories and model them separately.Extensive experiments on the realworld drug datasets,RTHNE_DTI produces high efficiency and outperforms other state-of-the-art methods.RTHNE_DTI can be further used to predict the interaction between unknown interaction drug-target pairs.

Electron-Elastic-Wave Interaction in a Two-Dimensional Topological Insulator

The interaction between an electron and an elastic wave is investigated for HgTe and InAs-GaSb quantum wells. The well-known Bernevig Hughes-Zhang model, i.e., the 4 × 4 model for a two-dimensional （2D） topological insulator （TI）, is extended to include the terms that describe the coupling between the electron and the elastic wave. The influence of this interaction on the transport properties of the 2DTI and of the edge states is discussed. As the electron-like and hole-like carriers interact with the elastic wave differently due to the crystal symmetry of the 2DTI, one may utilize the elastic wave to tune^control the transport property of charge carriers in the 2DTI. The extended 2DTI model also provides the possibility to investigate the backscattering of edge states of a 2DTI more realistically.

预测药物-靶点相互作用的异构网络嵌入模型研究

药物-靶点相互作用(DTIs)预测是药物发现的重要过程。随着计算技术的发展,基于生物数据的计算方法正高效率地加速这一过程。然而,这些方法大多忽略了靶标和药物的序列特征和异构性。本研究通过机器学习方法,提出多重网络嵌入框架(MLB-NEDTP)的预测模型,首先分析序列特征,然后将融合的特征嵌入多层异质信息网络中,以提高预测性能。多个数据集训练和验证结果表明,该模型与ATOMNET等最新模型相比具有明显优势。

Collaborative Matrix Factorization with Soft Regularization for Drug-Target Interaction Prediction

Identifying the potential drug-target interactions(DTI)is critical in drug discovery.The drug-target interaction prediction methods based on collaborative filtering have demonstrated attractive prediction performance.However,many corresponding models cannot accurately express the relationship between similarity features and DTI features.In order to rationally represent the correlation,we propose a novel matrix factorization method,so-called collaborative matrix factorization with soft regularization(SRCMF).SRCMF improves the prediction performance by combining the drug and the target similarity information with matrix factorization.In contrast to general collaborative matrix factorization,the fundamental idea of SRCMF is to make the similarity features and the potential features of DTI approximate,not identical.Specifically,SRCMF obtains low-rank feature representations of drug similarity and target similarity,and then uses a soft regularization term to constrain the approximation between drug(target)similarity features and drug(target)potential features of DTI.To comprehensively evaluate the prediction performance of SRCMF,we conduct cross-validation experiments under three different settings.In terms of the area under the precision-recall curve(AUPR),SRCMF achieves better prediction results than six state-of-the-art methods.Besides,under different noise levels of similarity data,the prediction performance of SRCMF is much better than that of collaborative matrix factorization.In conclusion,SRCMF is robust leading to performance improvement in drug-target interaction prediction.

DeepPurpose-based drug discovery in chondrosarcoma

Background:Chondrosarcoma(CS)is the second most common primary bone tumor,accounting for approximately30%of all malignant bone tumors.Unfortunately,the efficacy of currently available drug therapies is limited.Therefore,this study aimed to explore drug therapies for CS using novel computational methods.Methods:In this study,text mining,Gene Codis STRING,and Cytoscape were used to identify genes closely related to CS,and the Drug Gene Interaction Database(DGIdb)was used to select drugs targeting the genes.Drug-target interaction prediction was performed using Deep Purpose,to finally obtain candidate drugs with the highest predicted binding affinities.Results:Text-mining searches identified 168 genes related to CS.Gene enrichment and protein-protein interaction analysis generated 14 genes representing 10 pathways using Gene Codis,STRING,and Cytoscape.Seventy drugs targeting genes closely related to CS were analyzed using DGIdb.Deep Purpose recommended 25 drugs,including integrin beta 3 inhibitors,hypoxia-inducible factor 1 alpha inhibitors,E1A binding protein P300 inhibitors,vascular endothelial growth factor A inhibitors,AKT1 inhibitors,tumor necrosis factor inhibitors,transforming growth factor beta 1 inhibitors,interleukin 6 inhibitors,mitogen-activated protein kinase 1 inhibitors,and protein tyrosine kinase inhibitors.Conclusion:Drug discovery using in silico text mining and Deep Purpose may be an effective method to explore drugs targeting genes related to CS.

Identifying drug-target proteins based on network features

Proteins rarely function in isolation inside and outside cells, but operate as part of a highly intercon- nected cellular network called the interaction network. Therefore, the analysis of the properties of drug-target proteins in the biological network is especially helpful for understanding the mechanism of drug action in terms of informatics. At present, no detailed characterization and description of the topological features of drug-target proteins have been available in the human protein-protein interac- tion network. In this work, by mapping the drug-targets in DrugBank onto the interaction network of human proteins, five topological indices of drug-targets were analyzed and compared with those of the whole protein interactome set and the non-drug-target set. The experimental results showed that drug-target proteins have higher connectivity and quicker communication with each other in the PPI network. Based on these features, all proteins in the interaction network were ranked. The results showed that, of the top 100 proteins, 48 are covered by DrugBank; of the remaining 52 proteins, 9 are drug-target proteins covered by the TTD, Matador and other databases, while others have been dem- onstrated to be drug-target proteins in the literature.