Prediction of Potential Disease-Associated MicroRNAs by Using Neural Networks.

Identifying disease-related microRNAs (miRNAs) is an essential but challenging task in bioinformatics research. Much effort has been devoted to discovering the underlying associations between miRNAs and diseases. However, most studies mainly focus on designing advanced methods to improve prediction accuracy while neglecting to investigate the link predictability of the relationships between miRNAs and diseases. In this work, we construct a heterogeneous network by integrating neighborhood information in the neural network to predict potential associations between miRNAs and diseases, which also consider the imbalance of datasets. We also employ a new computational method called a neural network model for miRNA-disease association prediction (NNMDA). This model predicts miRNA-disease associations by integrating multiple biological data resources. Comparison of our work with other algorithms reveals the reliable performance of NNMDA. Its average AUC score was 0.937 over 15 diseases in a 5-fold cross-validation and AUC of 0.8439 based on leave-one-out cross-validation. The results indicate that NNMDA could be used in evaluating the accuracy of miRNA-disease associations. Moreover, NNMDA was applied to two common human diseases in two types of case studies. In the first type, 26 out of the top 30 predicted miRNAs of lung neoplasms were confirmed by the experiments. In the second type of case study for new diseases without any known miRNAs related to it, we selected breast neoplasms as the test example by hiding the association information between the miRNAs and this disease. The results verified 50 out of the top 50 predicted breast-neoplasm-related miRNAs.

Prediction of Drug-Gene Interaction by Using Metapath2vec.

Heterogeneous information networks (HINs) currently play an important role in daily life. HINs are applied in many fields, such as science research, e-commerce, recommendation systems, and bioinformatics. Particularly, HINs have been used in biomedical research. Algorithms have been proposed to calculate the correlations between drugs and targets and between diseases and genes. Recently, the interaction between drugs and human genes has become an important subject in the research on drug efficacy and human genomics. In previous studies, numerous prediction methods using machine learning and statistical prediction models were proposed to explore this interaction on the biological network. In the current work, we introduce a representation learning method into the biological heterogeneous network and use the representation learning models metapath2vec and metapath2vec++ on our dataset. We combine the adverse drug reaction (ADR) data in the drug-gene network with causal relationship between drugs and ADRs. This article first presents an analysis of the importance of predicting drug-gene relationships and discusses the existing prediction methods. Second, the skip-gram model commonly used in representation learning for natural language processing tasks is explained. Third, the metapath2vec and metapath2vec++ models for the example of drug-gene-ADR network are described. Next, the kernelized Bayesian matrix factorization algorithm is used to complete the prediction. Finally, the experimental results of both models are compared with Katz, CATAPULT, and matrix factorization, the prediction visualized using the receiver operating characteristic curves are presented, and the area under the receiver operating characteristic values for three varying algorithm parameters are calculated.