Drug response prediction using biological signaling pathway-centric analysis of genetic variation

Abstract

Over the last few decades, remarkable progress has been made in cancer research and cancer treatment. More recently, cancer research has been even more accelerated by development of state-of-the-art experimental technologies including massively parallel sequencing technology and the importance of understanding the diversity of cancer is emerged rapidly. One of the biggest challenges for realizing precision medicine for cancer is understanding how different molecular landscapes of cancer lead to different responses to anticancer drugs. Many studies have analyzed the comprehensive anticancer drug-response profiles and genomic profiles of cancer cell lines to identify the relationship between the anticancer drug response and genomic alterations. However, few studies have focused on interpreting these profiles with a network perspective. In this work, we analyzed genomic alterations in cancer cell lines by considering which interac- tions in the signaling pathway are perturbed by genomic variations. With our interaction-centric approach, we identified drug response associated(DRA) interactions for 80 drugs including novel DRA interactions for two drugs (afatinib and ixabepilone) for which no gene-centric association could be found. The DRA interactions were feasible to interpret as further research as the genes consisting the identified DRA interactions were more enriched with known chemrsensi- tivity related genes and drug targets than the genes identified by gene-centric approaches. The identified DRA interactions were validated with other cancer pharmacogenomic data. When we compared the performance of classifiers for predicting the responses to 164 drugs, the classifiers trained with interaction-centric features outperformed the classifiers trained with gene-centric features, despite the smaller number of features. By incorporating the interaction information from the signaling pathways, we revealed asso- ciations between genomic alterations and drug responses that could be missed when using a gene-centric approach and showed that the interaction-centric information can be useful for predicting anti-cancer drug sensitivity of a cancer cell.