Research On The Relationship Between LncRNA And Disease Based On Random Walk

LncRNA is a ncRNA with a length of more than 200 nucleotides.It plays an irreplaceable role in many important biological processes such as cell proliferation,cell differentiation,epigenetic regulation,genome imprinting,and chromosome modification.lncRNA is closely related to human diseases,but the exact mechanism behind it is unclear.The importance of lncRNA research for understanding cell biology and disease mechanisms is obvious.In order to promote the positive development of research,it is necessary to discover new lncRNA-disease associations.However,most of the relationship between lncRNA and disease has yet to be confirmed,because biological experiments are time consuming and laborio.Therefore,many researchers have begun to study advanced and effective calculation methods to verify the relationships between lncRNAs and diseases.It provides a direction for the effective advancement of biological experiments,so as to saving manpower and material resources.In order to obtain more accurate similarity information and construct a more effective network model,this paper proposed a method called lncRNA-disease association prediction based on linear neighborhood similarity and unbalanced bi-random walk(LDA-LNSUBRW).The main contents are as follows:(1)Research related biological databases and download the data sets needed to establish this model from the database,including lncRNA-disease association matrix,lncRNA functional similarity matrix and disease semantic similarity matrix.First,the weighted K-nearest neighbor algorithm is used to preprocess the data set to solve the lack of model prediction performance caused by the sparsity of the data set.Secondly,use the linear neighborhood similarity method to construct a new lncRNA similarity matrix and disease similarity matrix,so as to obtain more accurate similarity information.Finally,based on the differences in the topology of lncRNA and disease network,the distance between two nodes in their respective subnetworks should be controlled by walking different steps.Therefore,this paper proposes an unbalanced bi-random walk algorithm to supplement the potential mapping relationship between lncRNA network and disease network.(2)In order to evaluate the performance of the model LDA-LNSUBRW,the leave-one-out cross-validation and five-fold cross-validation were used for performance evaluation,and the AUC values obtained are 0.8874 and 0.8632±0.0051,respectively.In addition,in order to further verify the accuracy of the prediction results,a case analysis was carried out for the lncRNA-disease association pairs related to lung cancer,breast cancer and osteosarcoma in the prediction results.Verification from the two aspects of whether there is record information in the database and whether there is document verification information,the results show that the accuracy of the prediction method in this paper has reached an ideal result in the three diseases,and the prediction result is relatively reliable.(3)This paper also proposes the direction of model improvement,using the sequential double random walk algorithm to replace the unbalanced double random walk algorithm based on the linear neighborhood similarity and the unbalanced double random walk prediction model,so as to obtain the new model LDA-LNSSXRW.Using leave-one-out cross-validation and five-fold cross-validation to evaluate the performance of the new model,the AUC values obtained are 0.8819 and 0.8455±0.0053,respectively.Although the effect is slightly lower than the model LDA-LNSUBRW,as the future integrates more biological information and builds a larger heterogeneous network,under specific biological computing problems,LDA-LNSSXRW may perform better.