Potential LncRNA-Disease Associations Prediction Based On Bidirectional Mass Diffusion

Increasing evidence has demonstrated that long non-coding RNAs(lncRNAs)play essential roles in various human complex diseases.Compared with protein-coding genes,the associations between diseases and lncRNAs are still not well studied.Hence,inferring disease-associated lncRNAs on a genome-wide scale helps us to understand the occurrence process of diseases,and prevent,diagnose,and treat diseases at the molecular level,which has great research and medical value.The associations between lncRNAs and diseases can be accurately identified by using traditional biological experimental methods,but such methods require high experimental conditions,labor-intensive,and time-consuming.In recent years,researchers in the field of computer are paying more attention to the development of bioinformatics,and are committed to applying intelligent algorithms such as statistics and artificial intelligence to the field of bioinformatics.Therefore,many bioinformatics computational models have been used to predict the lncRNA-disease association to reduce the cost of human and material resources.Utilizing existing data to predict potential lncRNA-disease associations can provide useful help for further research on lncRNA.Based on previous studies,this paper proposed a method called Potential lncRNA-disease associations prediction based on mass diffusion(LDAPMD).Firstly,this model collected the known associations between human lncRNAs and diseases.Secondly,the similarity information of lncRNA-lncRNA Gaussian nuclear similarity,disease-disease Gaussian nuclear similarity,and other similar information was obtained through data preprocessing,and this biological information was deployed into the network.Based on the principle of material diffusion recommendation algorithm,The model can realize the forward and backward propagation of weighted resources,standardize the data,and improve the prediction performance of the model.At the same time,the model reduces the iterative process of the whole computing network of existing methods,greatly reducing the model computational intensity.To verify the accuracy of the prediction results,we performed a case study by using the predicted result of the lncRNA-disease association which related to colorectal cancer,lung cancer,and breast cancer.These results were tested from the public database and the validation information of the existing literature.The results show that the accuracy of this prediction method in all three diseases is greater than 90%,and the prediction is reliable.After leave-one-out cross-validation(LOOCV)and 5-fold cross-validation,the AUC value of LDAPMD reached 0.9207 and 0.9371,respectively.This method that Potential lncRNA-disease associations prediction based on mass diffusion has good performance in both evaluation indicators and case studies,and has strong robustness.LDAPMD has important significance for the lncRNA-disease association prediction study.