Research On Disease-related MiRNA Prediction Method Based On Phenotype-miRNA Network

MiRNAs are a class of short noncoding RNAs typically about 22?25 nucleotides in length.They regulate gene expression in the posttranscriptional level and participate in many important life processes,including cell division,proliferation,apoptosis and so on.In recent years,more and more studies have shown that miRNAs play important roles in the emergence and development of many diseases.Although a large number of miRNA-relevant data has been found and accumulated,only a few of miRNA-disease associations have been manifested due to the limitation of experimental technique.Therefore,it is important to develop bio informatics methods to uncover miRNA-disease association.Currently,there are some problems exist in disease-related miRNA prediction methods,such as the low prediction accuracy,the new miRNA-disease association prediction and so on.To deal with these problems,two disease-related miRNA prediction methods based on the phenotypic-miRN A network are proposed in this paper.The specific work is as follows.1.A new disease-ralated miRNA prediction method based on the global network similarity(NetGS)is proposed.First,this method obtains the global network similarities by calculating the graph Laplacian scores.Then miRNA-disease association scores are obtained by the randow walk algorithm and Pearson correlation formula.This method not only has advantage than existing methods on prediction accuracy,but also can be used to predict new miRNA-disease association.2.A new diffusion-based disease-related miRNA prediction method is proposed(NDBM).In this paper,we put forword the idea that extracting miRNA similarity and disease similarity from known disease-miRNA associations for the first time.Based on it,two similarity networks are constructed by fusing multiple information(miRNA similarity network and disease similarity net work).This method frist predict disease-related miRNAs on the miRNA similarity network and disease similarity network respectively.Then,the optimizing formula is used to combine the predicted results.The experimental results show that NDBM has improved the prediction accuarcy.Besides,NDBM and NetGS are analyzed and compared in this paper.The results show that each method has its own advantages.NetGS is suitable for new miRNA-disease association prediction and NDBM has better prediction performance in generally miRNA-disease association prediction.