A heterogeneous graph neural network with automatic discovery of effective metapaths for drug-target interaction prediction

Menglong Zhang, Yue Hong, Lian Shen, Shiyu Xu, Yanni Xu, Xinyi Zhang, Juan Liu, Xiangrong Liu

Abstract: Highlights•ADEM clusters entities through a triplet loss function to obtain meaningful features.•ADEM can automatically discover the most useful metapaths at the fine-grained level.•We establish a DTP-graph to predict DTI and use MixHop to process it.•The results demonstrate that ADEM outperforms other state-of-the-art approaches.