Fig. 1

The framework of DeepAEG. DeepAEG mainly includes three parts: data enhancement, drug feature mining and gene expression feature extraction. The drug feature mining module is divided into CGCN module and transformer module. SMILES expression of cancer drugs is data-enhanced after input and converted into advanced latent characterization by the Encoder and Transformer sections of the GCNC, respectively. The features of cancer cell profiles extracted by the sub-networks are connected and input into the classifier network