A Model Based on Random Walk with Restart to Predict CircRNA-Disease Associations on Heterogeneous Network

Abstract

Recent studies show that circRNAs have critical roles in many biological processes. Knowing the associations between circRNAs and diseases may contribute to the understanding of the mechanism of circRNAs and to the diagnostic and therapeutic methods of diseases at the molecular level. A small number of computation models have been developed to estimate CircRNA-disease associations. Therefore, in this study, a computational model has been developed. Similarity matrices have been obtained for circRNA and disease respectively by applying gaussian on the data obtained from the circRNADisease database. Then, random walk with restart algorithm applied on the combined matrices. The AUC value was obtained by 5-fold cross validation is 0.861 and this demonstrates the reliability of the model.