Fig. 1
From: Drug discovery and mechanism prediction with explainable graph neural networks
The architecture of the proposed model XGDP for drug response and mechanism prediction. Molecular graph, node features and edge features are extracted from the drug molecule, and GNN is used for learning the latent features of drugs. CNN is applied to compress the gene expression features from cancer cell lines. Then two multi-head cross-attention layers are leveraged to combine drug and cell features, and the drug response is predicted with the integrated features.
