Prediction Of LncRNA-disease Association Based On Internetwork Random Walk Algorithm

The expression of long non-coding RNA is closely related to the pathogenesis of many diseases.They may be important biomarkers for the diagnosis,prognosis and treatment of these diseases.Extrapolating the potential lncRNA-disease association is essential to reveal the pathogenesis of the disease,develop new drugs,and optimize personalized treatment.However,biological experiments are time-consuming and expensive to verify lncRNA-disease correlation.The effective computational models to predict lncRNA-disease correlation has become an important research topic in computational biology.Based on the internetwork random walk algorithm,this paper proposes a new network-based method to predict the potential lncRNA-disease association.Firstly,disease semantic similarity,lncRNA functional similarity and gene functional similarity were calculated to construct disease similarity network,lncRNA similarity network and gene similarity network.Then,lncRNA-disease association,lncRNA-gene association and gene-disease association were incorporated into the three networks.Finally,several random walk steps were performed in the three networks to infer the relationship between diseases and lncRNA in the corresponding networks.The model was applied to the association information data of disease-lncRNA,lncRNA-gene and disease-gene in LncRNADisease2.0 database to predict their association information.Finally,the prediction effect of the model was evaluated by using 10 fold cross validation and the areas under the receiver operating characteristic curve.The results show that compared with other algorithms,this algorithm has better AUC under 10 folds cross validation,which show that the algorithm has a better prediction performance.Furthermore,the lncRNA related to gastric cancer and colorectal cancer was predicted by using the algorithm.The predicted results are consistent with the known medical data,which shows that the algorithm is effective in predicting new disease-related lnRNA.In addition,this algorithm is also applied to the prediction of gene-disease association,and the results obtained are compared with the results of other existing models.The results imply that this model can improve the prediction effect of gene-disease association.