BPLLDA: Predicting lncRNA-Disease Associations Based on Simple Paths With Limited Lengths in a Heterogeneous Network

In recent years, it has been increasingly clear that long non-coding RNAs (lncRNAs) play critical roles in many basic and vital biological processes associated with various human diseases. However, the underlying mechanisms are yet to be discovered. Inferring potential lncRNA-disease associations is essential to revealing the secrets behind diseases, developing novel drugs, and conducting personalized treatment. However, biological experiments in validating lncRNA-disease associations are very time-consuming and costly, which urge the need for developing effective computational models. In this study, we proposed a method called BPLLDA to predict lncRNA-disease associations based on paths of fixed lengths in a heterogeneous lncRNA-disease associations network. Specifically, BPLLDA first constructs a heterogeneous lncRNA-disease network by integrating known lncRNA-disease associations, disease semantic similarity, lncRNA functional similarity and Gaussian interaction profile kernel similarity for lncRNAs and diseases. It then infers the probability of an lncRNA-disease association based on paths of limited lengths connecting them in the heterogeneous network. Comparing to existing methods, BPLLDA has a few benefits including not demanding negative samples and the ability to predict novel lncRNAs (without known associated disease) and novel diseases (without known associated lncRNA). BPLLDA was applied to a canonical lncRNA-disease associations database LncRNADisease together with two popular computational methods LRLSLDA and GrwLDA. The leave-one-out cross-validation (LOOCV) AUCs of BPLLDA are 0.87117, 0.82403, and 0.78528 respectively for overall, novel lncRNA, and novel disease prediction, outperforming the two compared methods. In addition, cervical cancer, glioma, and non-small cell lung cancer were selected as case studies, for which the top 5 lncRNA-disease associations were verified by newly updated LncRNADisease database and recently published literatures. As an indication, BPLLDA will contribute to the understanding of complex diseases by inferring potential lncRNA-disease associations for further experimental validation.