| MicroRNAs(mi RNAs)are a class of small,endogenous,singlestranded,noncoding RNAs about 20-24 nucleotides that mainly repress the expression of target m RNAs at the posttranscriptional level by binding to the 3′-untranslated region of target mRNAs through sequence-specific base pairing,resulting in target mRNAs cleavage or translation inhibition.More and more studies have shown that mi RNAs played an important role in all kinds of biological processes.The mutations and dysfunctions of mi RNAs could lead to a variety of diseases.Thus,identifying regulatory relationships between mi RNAs and diseases is critical,which has become a hot topic in recent years.The early study used biological experiments to seek the effect of a single factor on experiments,for the sake of obtaining more accurated results.However,since biological experiments is time consuming and high cost,researchers developed more efficient calculation methods to solve this problem.Therefore,development of bioinformatics to explore the relationship of mi RNAs-diseases is important.Current computational methods can be mainly divided into two categories:(1)methods based on similarity measurement,(2)methods based on machine learning.The former approaches predict mi RNA-disease associations by measuring similarity of nodes in the biological networks but they need to build high quality biological networks.The latter approaches apply machine learning algorithms to this problem but these methods need to build a negative collection of high credibility.Based on the shortcomings above,this paper proposes two novel calculation models which are Thr RW and BNPDCMDA to predict mi RNA-diseases associations.The former employs genes into this problem,making full use of the similarity of genes,mi RNAs and diseases as well as the associations among them,to construct an unbalanced random walk model to infer mi RNA-disease associations.The latter firstly performs density clustering based on the similarity of mi RNAs,then constructs mi RNA-disease double-layer combining diseases and results of density clustering mi RNAs,after that applies bipartite network projection to the network above,finally accomplishes the predictions for mi RNA-disease associations.To validate the effectiveness of the two proposed calculation models,this paper employs the leave-one-out cross-validation method to compare with other effective methods.The results show that the AUC values of BNPDCMDA and Thr RW can achieve 86.24% and 99.08%,respectively,which performs better than other state-ofthe-art methods.In addition,we also predict associated mi RNAs for some common diseases(such as breast cancer and lung cancer).The results show that most of the associations predicted by the two models are supported by literature,which further indicates the effectiveness of Thr RW and BNPDCMDA. |
