A novel semi-supervised model for miRNA-disease association prediction based on -norm graph

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• MicroRNAs (miRNAs) play crucial roles in biological and pathological processes.
• Identifying miRNA-disease associations is vital for understanding disease mechanisms.
• Experimental methods for miRNA-disease association discovery are time-consuming and labor-intensive.

Purpose of the Study:

• To propose a novel semi-supervised model for predicting miRNA-disease associations.
• To enhance the efficiency and accuracy of computational methods for miRNA-disease association discovery.

Main Methods:

• A semi-supervised model utilizing a novel graph-based approach.
• Recalculation of miRNA functional similarities and disease semantic similarities using MeSH descriptors and HMDD.
• Iterative updates of similarity and association matrices in both miRNA and disease spaces.

Main Results:

• The proposed method achieved high performance with AUCs of 0.943 and 0.946.
• Case studies confirmed a significant portion of predicted miRNA-disease associations in five common human diseases.
• Results suggest that miRNAs within the same family or cluster may function together in diseases.

Conclusions:

• The developed method is a reliable and efficient tool for predicting miRNA-disease associations.
• The findings highlight the utility of computational approaches in accelerating miRNA-disease association discovery.
• The study provides valuable insights into the functional roles of miRNAs in human diseases.