HOPEXGB: A Consensual Model for Predicting miRNA/lncRNA-Disease Associations Using a Heterogeneous Disease-miRNA-lncRNA Information Network

Predicting disease-related microRNAs(miRNAs) and longnoncodingRNAs (lncRNAs) is crucial to find new biomarkers for the prevention,diagnosis, and treatment of complex human diseases. Computationalpredictions for miRNA/lncRNA-disease associations are of great practicalsignificance, since traditional experimental detection is expensiveand time-consuming. In this paper, we proposed a consensual machine-learningtechnique-based prediction approach to identify disease-related miRNAsand lncRNAs by high-order proximity preserved embedding (HOPE) andeXtreme Gradient Boosting (XGB), named HOPEXGB. By connecting lncRNA,miRNA, and disease nodes based on their correlations and relationships,we first created a heterogeneous disease-miRNA-lncRNA (DML) informationnetwork to achieve an effective fusion of information on similarities,correlations, and interactions among miRNAs, lncRNAs, and diseases.In addition, a more rational negative data set was generated basedon the similarities of unknown associations with the known ones, soas to effectively reduce the false negative rate in the data set formodel construction. By 10-fold cross-validation, HOPE shows betterperformance than other graph embedding methods. The final consensualHOPEXGB model yields robust performance with a mean prediction accuracyof 0.9569 and also demonstrates high sensitivity and specificity advantagescompared to lncRNA/miRNA-specific predictions. Moreover, it is superiorto other existing methods and gives promising performance on the externaltesting data, indicating that integrating the information on lncRNA-miRNAinteractions and the similarities of lncRNAs/miRNAs is beneficialfor improving the prediction performance of the model. Finally, casestudies on lung, stomach, and breast cancers indicate that HOPEXGBcould be a powerful tool for preclinical biomarker detection and bioexperimentpreliminary screening for the diagnosis and prognosis of cancers.HOPEXGB is publicly available at https://github.com/airpamper/HOPEXGB.

Automated summary

This study proposes a prediction approach called HOPEXGB to identify disease-related miRNAs and lncRNAs. By constructing a heterogeneous information network and using machine learning techniques, it achieves accurate predictions and demonstrates superiority over existing methods.