Abstract:There is a growing body of evidence from biological experiments suggesting that microRNAs (miRNAs) play a significant regulatory role in both diverse cellular activities and pathological processes. Exploring miRNA-disease associations not only can decipher pathogenic mechanisms but also provide treatment solutions for diseases. As it is inefficient to identify undiscovered relationships between diseases and miRNAs using biotechnology, an explosion of computational methods have been advanced. However, the prediction accuracy of existing models is hampered by the sparsity of known association network and single-category feature, which is hard to model the complicated relationships between diseases and miRNAs.

Predicting miRNA-disease associations based on graph attention network with multi-source information