Higher-Order Proximity-Based MiRNA-Disease Associations Prediction

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• MicroRNA (miRNA)-disease association prediction is crucial for understanding disease pathogenesis and developing personalized diagnostics.
• Existing methods often overlook implicit higher-order proximity among miRNAs and diseases, limiting prediction accuracy.
• Explicit information like miRNA functional similarity and disease semantic similarity has been utilized, but higher-order implicit relationships remain underexplored.

Purpose of the Study:

• To propose a novel computational approach, HOP_MDA (Higher-Order Proximity based MiRNA and Disease Association Prediction), for predicting potential miRNA-disease associations.
• To effectively integrate both explicit interaction data and implicit higher-order proximity information between miRNAs and diseases.
• To develop an efficient prediction technique, HOPA_MDA, for faster and accurate association predictions.

Main Methods:

• Developed HOP_MDA, encoding explicit interactions and implicit higher-order proximity into a parameterized prediction matrix.
• Employed a supervised learning approach to optimize weight parameters based on known miRNA-disease associations.
• Introduced HOPA_MDA, a higher-order proximity approximation technique for enhanced prediction efficiency.

Main Results:

• The proposed HOPA_MDA method achieved high average AUC values of 0.921+/-0.002 and 0.944+/-0.0015 on two real datasets.
• Demonstrated the capability of the method to predict potential miRNAs for novel diseases lacking prior miRNA association data.
• Validated the effectiveness of incorporating higher-order proximity in miRNA-disease association prediction.

Conclusions:

• HOP_MDA and its approximation HOPA_MDA offer a significant advancement in miRNA-disease association prediction by leveraging higher-order proximity.
• The method provides a robust framework for identifying disease-related miRNAs, aiding in diagnostics and pathogenesis research.
• This approach holds promise for discovering novel miRNA-disease links, particularly for diseases with limited existing information.