Fig. 3

Analysis of features extracted by local and global graph learning. a shows the t-SNE visualization of features extracted by iPiDA-A, iPiDA-G, and iPiDA-L, respectively. b and c show local context graphs for two example positive and two example negative piRNA-disease pairs in \({\mathbb{S}}_{\text{independent}}\), respectively. d shows the heatmap of features extracted by iPiDA-G and iPiDA-L for four example piRNA-disease pairs