Drug-drug similarity measure and its applications

Potential Target Discovery and Drug Repurposing for Coronaviruses: Study Involving a Knowledge Graph-Based Approach

BACKGROUND

The global pandemics of severe acute respiratory syndrome, Middle East respiratory syndrome, and COVID-19 have caused unprecedented crises for public health. Coronaviruses are constantly evolving, and it is unknown which new coronavirus will emerge and when the next coronavirus will sweep across the world. Knowledge graphs are expected to help discover the pathogenicity and transmission mechanism of viruses.

OBJECTIVE

The aim of this study was to discover potential targets and candidate drugs to repurpose for coronaviruses through a knowledge graph-based approach.

METHODS

We propose a computational and evidence-based knowledge discovery approach to identify potential targets and candidate drugs for coronaviruses from biomedical literature and well-known knowledge bases. To organize the semantic triples extracted automatically from biomedical literature, a semantic conversion model was designed. The literature knowledge was associated and integrated with existing drug and gene knowledge through semantic mapping, and the coronavirus knowledge graph (CovKG) was constructed. We adopted both the knowledge graph embedding model and the semantic reasoning mechanism to discover unrecorded mechanisms of drug action as well as potential targets and drug candidates. Furthermore, we have provided evidence-based support with a scoring and backtracking mechanism.

RESULTS

The constructed CovKG contains 17,369,620 triples, of which 641,195 were extracted from biomedical literature, covering 13,065 concept unique identifiers, 209 semantic types, and 97 semantic relations of the Unified Medical Language System. Through multi-source knowledge integration, 475 drugs and 262 targets were mapped to existing knowledge, and 41 new drug mechanisms of action were found by semantic reasoning, which were not recorded in the existing knowledge base. Among the knowledge graph embedding models, TransR outperformed others (mean reciprocal rank=0.2510, Hits@10=0.3505). A total of 33 potential targets and 18 drug candidates were identified for coronaviruses. Among them, 7 novel drugs (ie, quinine, nelfinavir, ivermectin, asunaprevir, tylophorine, Artemisia annua extract, and resveratrol) and 3 highly ranked targets (ie, angiotensin converting enzyme 2, transmembrane serine protease 2, and M protein) were further discussed.

CONCLUSIONS

We showed the effectiveness of a knowledge graph-based approach in potential target discovery and drug repurposing for coronaviruses. Our approach can be extended to other viruses or diseases for biomedical knowledge discovery and relevant applications.

Identification of Robust Antibiotic Subgroups by Integrating Multi-Species Drug-Drug Interactions

Previous studies have shown that antibiotics can be divided into groups, and drug-drug interactions (DDI) depend on their groups. However, these studies focused on a specific bacteria strain (i.e., Escherichia coli BW25113). Existing datasets often contain noise. Noisy labeled data may have a bad effect on the clustering results. To address this problem, we developed a multi-source information fusion method for integrating DDI information from multiple bacterial strains. Specifically, we calculated drug similarities based on the DDI network of each bacterial strain and then fused these drug similarity matrices to obtain a new fused similarity matrix. The fused similarity matrix was combined with the T-distributed stochastic neighbor embedding algorithm, and hierarchical clustering algorithm can effectively identify antibiotic subgroups. These antibiotic subgroups are strongly correlated with known antibiotic classifications, and group-group interactions are almost monochromatic. In summary, our method provides a promising framework for understanding the mechanism of action of antibiotics and exploring multi-species group-group interactions.

A Graph Representation Approach Based on Light Gradient Boosting Machine for Predicting Drug-Disease Associations

Determining the association between drug and disease is important in drug development. However, existing approaches for drug-disease associations (DDAs) prediction are too homogeneous in terms of feature extraction. Here, a novel graph representation approach based on light gradient boosting machine (GRLGB) is proposed for prediction of DDAs. After the introduction of the protein into a heterogeneous network, nodes features were extracted from two perspectives: network topology and biological knowledge. Finally, the GRLGB classifier was applied to predict potential DDAs. GRLGB achieved satisfactory results on Bdataset and Fdataset through 10-fold cross-validation. To further prove the reliability of the GRLGB, case studies involving anxiety disorders and clozapine were conducted. The results suggest that GRLGB can identify novel DDAs.

DTiGNN: Learning drug-target embedding from a heterogeneous biological network based on a two-level attention-based graph neural network

MOTIVATION

In vitro experiment-based drug-target interaction (DTI) exploration demands more human, financial and data resources. In silico approaches have been recommended for predicting DTIs to reduce time and cost. During the drug development process, one can analyze the therapeutic effect of the drug for a particular disease by identifying how the drug binds to the target for treating that disease. Hence, DTI plays a major role in drug discovery. Many computational methods have been developed for DTI prediction. However, the existing methods have limitations in terms of capturing the interactions via multiple semantics between drug and target nodes in a heterogeneous biological network (HBN).

METHODS

In this paper, we propose a DTiGNN framework for identifying unknown drug-target pairs. The DTiGNN first calculates the similarity between the drug and target from multiple perspectives. Then, the features of drugs and targets from each perspective are learned separately by using a novel method termed an information entropy-based random walk. Next, all of the learned features from different perspectives are integrated into a single drug and target similarity network by using a multi-view convolutional neural network. Using the integrated similarity networks, drug interactions, drug-disease associations, protein interactions and protein-disease association, the HBN is constructed. Next, a novel embedding algorithm called a meta-graph guided graph neural network is used to learn the embedding of drugs and targets. Then, a convolutional neural network is employed to infer new DTIs after balancing the sample using oversampling techniques.

RESULTS

The DTiGNN is applied to various datasets, and the result shows better performance in terms of the area under receiver operating characteristic curve (AUC) and area under precision-recall curve (AUPR), with scores of 0.98 and 0.99, respectively. There are 23,739 newly predicted DTI pairs in total.

On the similarity measures of N-cubic Pythagorean fuzzy sets using the overlapping ratio

The similarity measures are essential concepts to discuss the closeness between sets. Fuzzy similarity measures and intuitionistic fuzzy similarity measures dealt with the incomplete and inconsistent data more efficiently. With time in decision-making theory, a complex frame of the environment that occurs cannot be specified entirely by these sets. A generalization like the Pythagorean fuzzy set can handle such a situation more efficiently. The applicability of this set attracted the researchers to generalize it into N-Pythagorean, interval-valued N-Pythagorean, and N-cubic Pythagorean sets. For this purpose, first, we define the overlapping ratios of N-interval valued Pythagorean and N-Pythagorean fuzzy sets. In addition, we define similarity measures in these sets. We applied this proposed measure for comparison analysis of plagiarism software.

BioChemDDI: Predicting Drug-Drug Interactions by Fusing Biochemical and Structural Information through a Self-Attention Mechanism

Simple Summary

Throughout history, combining drugs has been a common method in the fight against complex diseases. However, potential drug–drug interactions could give rise to unknown toxicity issues, which requires the urgent proposal of efficient methods to identify potential interactions.We use computer technology and machine learning techniques to propose a novel computational framework to calculate scores of drug–drug interaction probability for simplifying the screening process. Additionally, we built an online prescreening tool for biological researchers to further verify possible interactions in the fields of biomedicine and pharmacology. Overall, our study can provide new insights and approaches for rapidly identifying potential drug–drug interactions.

Abstract

During the development of drug and clinical applications, due to the co-administration of different drugs that have a high risk of interfering with each other’s mechanisms of action, correctly identifying potential drug–drug interactions (DDIs) is important to avoid a reduction in drug therapeutic activities and serious injuries to the organism. Therefore, to explore potential DDIs, we develop a computational method of integrating multi-level information. Firstly, the information of chemical sequence is fully captured by the Natural Language Processing (NLP) algorithm, and multiple biological function similarity information is fused by Similarity Network Fusion (SNF). Secondly, we extract deep network structure information through Hierarchical Representation Learning for Networks (HARP). Then, a highly representative comprehensive feature descriptor is constructed through the self-attention module that efficiently integrates biochemical and network features. Finally, a deep neural network (DNN) is employed to generate the prediction results. Contrasted with the previous supervision model, BioChemDDI innovatively introduced graph collapse for extracting a network structure and utilized the biochemical information during the pre-training process. The prediction results of the benchmark dataset indicate that BioChemDDI outperforms other existing models. Moreover, the case studies related to three cancer diseases, including breast cancer, hepatocellular carcinoma and malignancies, were analyzed using BioChemDDI. As a result, 24, 18 and 20 out of the top 30 predicted cancer-related drugs were confirmed by the databases. These experimental results demonstrate that BioChemDDI is a useful model to predict DDIs and can provide reliable candidates for biological experiments. The web server of BioChemDDI predictor is freely available to conduct further studies.

A Novel COVID-19-Related Drug Discovery Approach Based on Non-Equidimensional Data Clustering

The novel coronavirus disease (COVID-19) caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has spread all over the world. Since currently no effective antiviral treatment is available and those original inhibitors have no significant effect, the demand for the discovery of potential novel SARS-CoV-2 inhibitors has become more and more urgent. In view of the availability of the inhibitor-bound SARS-CoV-2 Mpro and PLpro crystal structure and a large amount of proteomics knowledge, we attempted using the existing coronavirus inhibitors to synthesize new ones, which combined the advantages of similar effective substructures for COVID-19 treatment. To achieve this, we first formulated this issue as a non-equidimensional inhibitor clustering and a following cluster center generating problem, where three essential challenges were carefully addressed, which are 1) how to define the distance between pairwise inhibitors with non-equidimensional molecular structure; 2) how to group inhibitors into clusters when the dimension is different; 3) how to generate the cluster center under this non-equidimensional condition. To be more specific, a novel matrix Kronecker product (p, m)-norm ⋅pm⊗ was first defined to induce the distance D_p(A, B) between two inhibitors. Then, the hierarchical clustering approach was conducted to find similar inhibitors, and a novel iterative algorithm–based Kronecker product (p, m)-norm was designed to generate individual cluster centers as the drug candidates. Numerical experiments showed that the proposed methods can find novel drug candidates efficiently for COVID-19, which has provided valuable predictions for further biological evaluations.

HINGRL: predicting drug-disease associations with graph representation learning on heterogeneous information networks

Identifying new indications for drugs plays an essential role at many phases of drug research and development. Computational methods are regarded as an effective way to associate drugs with new indications. However, most of them complete their tasks by constructing a variety of heterogeneous networks without considering the biological knowledge of drugs and diseases, which are believed to be useful for improving the accuracy of drug repositioning. To this end, a novel heterogeneous information network (HIN) based model, namely HINGRL, is proposed to precisely identify new indications for drugs based on graph representation learning techniques. More specifically, HINGRL first constructs a HIN by integrating drug-disease, drug-protein and protein-disease biological networks with the biological knowledge of drugs and diseases. Then, different representation strategies are applied to learn the features of nodes in the HIN from the topological and biological perspectives. Finally, HINGRL adopts a Random Forest classifier to predict unknown drug-disease associations based on the integrated features of drugs and diseases obtained in the previous step. Experimental results demonstrate that HINGRL achieves the best performance on two real datasets when compared with state-of-the-art models. Besides, our case studies indicate that the simultaneous consideration of network topology and biological knowledge of drugs and diseases allows HINGRL to precisely predict drug-disease associations from a more comprehensive perspective. The promising performance of HINGRL also reveals that the utilization of rich heterogeneous information provides an alternative view for HINGRL to identify novel drug-disease associations especially for new diseases.

Cascade Deep Forest With Heterogeneous Similarity Measures for Drug-Target Interaction Prediction

Drug repositioning is a method of systematically identifying potential molecular targets that known drugs may act on. Compared with traditional methods, drug repositioning has been extensively studied due to the development of multi-omics technology and system biology methods. Because of its biological network properties, it is possible to apply machine learning related algorithms for prediction. Based on various heterogeneous network model, this paper proposes a method named THNCDF for predicting drug-target interactions. Various heterogeneous networks are integrated to build a tripartite network, and similarity calculation methods are used to obtain similarity matrix. Then, the cascade deep forest method is used to make prediction. Results indicate that THNCDF outperforms the previously reported methods based on the 10-fold cross-validation on the benchmark data sets proposed by Y. Yamanishi. The area under Precision Recall curve (AUPR) value on the Enzyme, GPCR, Ion Channel, and Nuclear Receptor data sets is 0.988, 0.980, 0.938, and 0.906 separately. The experimental results well illustrate the feasibility of this method.

Heterogeneous graph inference with matrix completion for computational drug repositioning

MOTIVATION

Emerging evidence presents that traditional drug discovery experiment is time-consuming and high costs. Computational drug repositioning plays a critical role in saving time and resources for drug research and discovery. Therefore, developing more accurate and efficient approaches is imperative. Heterogeneous graph inference is a classical method in computational drug repositioning, which not only has high convergence precision, but also has fast convergence speed. However, the method has not fully considered the sparsity of heterogeneous association network. In addition, rough similarity measure can reduce the performance in identifying drug-associated indications.

RESULTS

In this article, we propose a heterogeneous graph inference with matrix completion (HGIMC) method to predict potential indications for approved and novel drugs. First, we use a bounded matrix completion (BMC) model to prefill a part of the missing entries in original drug-disease association matrix. This step can add more positive and formative drug-disease edges between drug network and disease network. Second, Gaussian radial basis function (GRB) is employed to improve the drug and disease similarities since the performance of heterogeneous graph inference more relies on similarity measures. Next, based on the updated drug-disease associations and new similarity measures of drug and disease, we construct a novel heterogeneous drug-disease network. Finally, HGIMC utilizes the heterogeneous network to infer the scores of unknown association pairs, and then recommend the promising indications for drugs. To evaluate the performance of our method, HGIMC is compared with five state-of-the-art approaches of drug repositioning in the 10-fold cross-validation and de novo tests. As the numerical results shown, HGIMC not only achieves a better prediction performance, but also has an excellent computation efficiency. In addition, cases studies also confirm the effectiveness of our method in practical application.

AVAILABILITY

The HGIMC software is freely available at https://github.com/BioinformaticsCSU/HGIMC, https://hub.docker.com/repository/docker/yangmy84/hgimc, and http://doi.org/10.5281/zenodo.4285640.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.