Incorporating higher order network structures to improve miRNA-disease association prediction based on functional...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• MicroRNAs (miRNAs) play crucial roles in biological processes, and their dysregulation is linked to complex diseases.
• Abnormal miRNA expression can serve as biomarkers for early disease detection and prognosis.
• Existing computational methods for predicting miRNA-disease associations (MDAs) often fail to fully exploit network patterns.

Purpose of the Study:

• To develop a novel computational model, HiSCMDA, for predicting miRNA-disease associations (MDAs).
• To improve MDA prediction accuracy by incorporating higher-order network structures.
• To enhance early disease detection and prognostic evaluation through better MDA prediction.

Main Methods:

• Integrated miRNA similarity, disease similarity, and MDA networks.
• Identified overlapping functional modules using higher-order connectivity patterns.
• Employed a path-based scoring function to infer potential MDAs within functional modules.

Main Results:

• HiSCMDA demonstrated superior performance across various datasets and evaluation metrics.
• The model achieved high accuracy in cross-validation and independent validation experiments.
• Case studies showed significant validation of predicted miRNA-neoplasm associations for colon and lung cancers.

Conclusions:

• Higher-order organizational structures in MDA networks offer valuable insights for prediction.
• HiSCMDA provides an effective approach for predicting miRNA-disease associations.
• The findings contribute to advancing early disease detection and prognostic strategies using miRNAs.