Predicting Drug-Disease Associations Based on Network Consistency Projection

With the increasing cost of traditional drug discovery, drug repositioning methods at low cost have attracting increasing attention. The generation of large amounts of biomedical data also provides unprecedented opportunities for drug repositioning research. However, how to effectively integrate different types of data is still a challenge for drug repositioning. In this paper, we propose a computational method using Network Consistency Projection for Drug-Disease Association (NCPDDA) prediction. First of all, our method proposes a new method for calculating one type of disease similarity. Moreover, since effective integration of data from multiple sources can improve prediction performance, the NCPDDA integrates multiple kinds of similarities. Then, considering that noise may affect the prediction performance of the model, the NCPDDA uses the similarity network fusion method to reduce the impact of noise. Finally, the network consistency projection is used to predict potential drug-disease associations. NCPDDA is compared with several classical drug repositioning methods, and the experimental results show that NCPDDA is superior to these methods. Moreover, the study of several representative drugs proves the practicality of NCPDDA in practical application.