Integrating multiple microRNA functional similarity networks for improved disease-microRNA association prediction

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• MicroRNAs (miRNAs) are crucial in disease pathogenesis, necessitating accurate identification for precision medicine.
• Existing methods for predicting miRNA-disease associations have limitations in capturing complex biological relationships.

Purpose of the Study:

• To develop and validate a novel computational method, multiplex-heterogeneous network for MiRNA-disease associations (MHMDA), for predicting disease-associated miRNAs.
• To leverage integrated miRNA functional similarity and disease similarity networks for enhanced prediction accuracy.

Main Methods:

• Constructed a multiplex-heterogeneous network integrating multiple miRNA functional similarity networks and a disease similarity network.
• Employed a tailored random walk with restart algorithm to predict miRNA-disease associations.
• Utilized experimentally validated and predicted miRNA-target interactions, alongside disease phenotypic similarities.

Main Results:

• MHMDA achieved high performance in leave-one-out and 5-fold cross-validation on human microRNA disease database and miR2Disease datasets.
• Achieved area under the receiver operating characteristic curve (AUC) values of 0.938 and 0.913, outperforming existing methods.
• Demonstrated robustness and stability across parameter variations and disease contexts.

Conclusions:

• The MHMDA method effectively predicts disease-miRNA associations by integrating diverse biological data.
• The multiplex-heterogeneous network approach enhances prediction accuracy, offering a robust tool for identifying novel disease-miRNA links.
• MHMDA shows significant potential for applications in precision medicine and understanding disease mechanisms.