CFMDA: collaborative filtering-based MiRNA-disease association prediction

MicroRNAs (miRNAs) are increasingly becoming the focus in a number of researches because abundant studies certify miRNAs play vital roles and have critical functions in various biologic processes. Considering the high cost of experiment research to miRNA-disease association, we explore the way to predict the miRNA-disease association using the extensive collaborative filtering in order to diagnose the diseases better. Specifically, we introduce the prediction model of collaborative filtering-based miRNA-disease association prediction (CFMDA) and verify the model by leave-one-out cross validation(LOOCV) and case validation. The CFMDA considers the miRNA functional similarity and disease similarity while uses minimal amount of related information. CFMDA achieves AUCs of 0.8364 using leave-one-out cross validation, which is the highest AUCs compared to other 5 methods. Meanwhile, we obtain more than 85% confirmation of predicted associations using three kinds of case validations. Generally, our method is faster and more effective than other state-of-the-art methods while it doesn’t need any negative samples.