BiG-DRP (and BiG-DRP+) are deep learning approaches for drug response prediction in cancer. These models leverage the chemical structure of drugs and the underlying relationships between drugs and cell lines through a bipartite graph and a heterogeneous graph convolutional network. They incorporate sensitive and resistant cell line information to form robust drug representations, significantly improving prediction performance. The tool is designed to enable prediction for unseen cell lines and can identify genes contributing significantly to model performance, pointing to important biological processes and signaling pathways. BiG-DRP is a valuable tool for pharmacogenomics studies and accelerating drug discovery.