Prediction of Potential Associations Between MicroRNA and Disease Based on Bayesian Probabilistic Matrix Factorization Model.

The association between microRNAs (miRNAs) and diseases is significant to understand the development and progression of many human diseases. Given the cost and complexity of biological experiments, the computational method for predicting the potential association between miRNAs and disease will be an effective complement. In this article, we have developed a model (microRNA and disease based on Bayesian probabilistic matrix factorization, MDBPMF) based on a fully Bayesian treatment of the probabilistic matrix factorization to find potential associations between miRNAs and diseases by using the HMDDv2.0 database, which contains proven miRNA-disease associations. We show that Bayesian probabilistic matrix factorization models can be efficiently trained using Markov chain Monte Carlo methods by applying them to the HMDDv2.0 database. MDBPMF achieves reliable prediction with an average area under receiver operating characteristic curve of 0.8755 for eight complex diseases based on fivefold cross-validation, which indeed outperforms the state-of-the-art method. In addition, a case study of lung cancer further verifies the utility of our method.