Looking at the BiG picture: incorporating bipartite graphs in drug response prediction

• DOI: 10.1093/bioinformatics/btac383
• Publication Date: 2022-06-08T12:00:38Z
• Authors: David Earl Hostal...; Yihui Li; Amin Emad
• Keywords: Drug response; Leverage (statistics); Pharmacogenomics; Python

The increasing number of publicly available databases containing drugs' chemical structures, their response in cell lines, and molecular profiles the lines has garnered attention to problem drug prediction. However, many existing methods do not fully leverage information that is shared among drugs with similar structure. As such, similarities terms line responses structures could prove be useful forming representations improve prediction accuracy.We present two deep learning approaches, BiG-DRP BiG-DRP+, for Our models take advantage structure underlying relationships through a bipartite graph heterogeneous convolutional network incorporate sensitive resistant representations. Evaluation our other state-of-the-art different scenarios shows incorporating this significantly improves performance. In addition, genes contribute performance also point important biological processes signaling pathways. Analysis predicted patients' tumors using model revealed associations between mutations sensitivity, illustrating utility pharmacogenomics studies.An implementation algorithms Python provided https://github.com/ddhostallero/BiG-DRP.Supplementary data are at Bioinformatics online.