TY  - JOUR
T1  - RGCNCDA: Relational graph convolutional network improves circRNA-disease association prediction by incorporating microRNAs
AU  - Chen, Yaojia
AU  - Wang, Yanpeng
AU  - Ding, Yijie
AU  - Su, Xi
AU  - Wang, Chunyu
JO  - Computers in Biology and Medicine
VL  - 143
SP  - 105322
PY  - 2022
DA  - 2022/04/01/
SN  - 0010-4825
DO  - https://doi.org/10.1016/j.compbiomed.2022.105322
UR  - https://www.sciencedirect.com/science/article/pii/S0010482522001147
KW  - CircRNA
KW  - Disease
KW  - circRNA-disease associations
KW  - microRNA
KW  - Relational graph convolution network
AB  - Recently, a large number of studies have indicated that circRNAs with covalently closed loops play important roles in biological processes and have potential as diagnostic biomarkers. Therefore, research on the circRNA-disease relationship is helpful in disease diagnosis and treatment. However, traditional biological verification methods require considerable labor and time costs. In this paper, we propose a new computational method (RGCNCDA) to predict circRNA-disease associations based on relational graph convolutional networks (R-GCNs). The method first integrates the circRNA similarity network, miRNA similarity network, disease similarity network and association networks among them to construct a global heterogeneous network. Then, it employs the random walk with restart (RWR) and principal component analysis (PCA) models to learn low-dimensional and high-order information from the global heterogeneous network as the topological features. Finally, a prediction model based on an R-GCN encoder and a DistMult decoder is built to predict the potential disease-associated circRNA. The predicted results demonstrate that RGCNCDA performs significantly better than the other six state-of-the-art methods in a 5-fold cross validation. Furthermore, the case study illustrates that RGCNCDA can effectively discover potential circRNA-disease associations.
ER  -