Predicting Disease-Associated N7-Methylguanosine (m⁷G) Sites via Random Walk on Heterogeneous Network

Recent studies revealed that the modification of N7-methylguanosine (m(7)G) has associations with many human diseases. Effectively identifying disease-associated m(7)G methylation sites would provide crucial clues for disease diagnosis and treatment. Previous studies have developed computational methods to predict disease-associated m(7)G sites based on similarities among m(7)G sites and diseases. However, few have focused on the influence of the known m(7)G-disease association information on calculating similarity measures of m(7)G site and disease, which potentially promotes the identification of the disease-associated m(7)G sites. In this work, we propose a computational method called m(7)GDP-RW to predict m(7)G-disease associations by random walk algorithm. m(7)GDP-RW first incorporates the feature information of m(7)G site and disease with the known m(7)G-disease associations to compute m(7)G site similarity and disease similarity. Then m(7)GDP-RW combines the known m(7)G-disease associations with the computed similarity of m(7)G site and disease to construct a m(7)G-disease heterogeneous network. Finally, m(7)GDP-RW utilizes a two-pass random walk with restart algorithm to find novel m(7)G-disease associations on the heterogeneous network. The experimental results show that our method achieves higher prediction accuracy compared to the existing methods. The study case also demonstrates the effectiveness of m(7)GDP-RW in discovering potential m(7)G-disease associations.