Predicting disease-gene associations through self-supervised mutual infomax graph convolution network

Jiancong Xie, Jiahua Rao, Junjie Xie, Huiying Zhao, Yuedong Yang

Abstract: Highlights•We constructed disease-disease and gene-gene association networks to facilitate disease-gene association prediction.•The inevitable noise from external data can affect the model performance.•We proposed MiGCN with a self-supervised infomax module to eliminate noise from the external data.•MiGCN achieved state-of-the-art performance compared with existing baselines.