| Cancer is a notoriously complex disease with multiple factors and biological processes involved in carcinogenesis. Monotherapy for cancer treatment in the clinic has been frequently associated with acquired resistance and side effects for years, whereas combinatorial drug therapy has attracted much more attention and was widely explored, especially the synergistic drugs. Synergistic drug combinations not only exhibit a greater overall therapeutic effect than the sum of the individual effects, but present largely reduced side effects because of the lower dosage of each ingredient compared with that used in monotherapy. Furthermore, there is a major obstacle, stems from high inter-tumor heterogeneity, to apply the traditional one-size-fits-all chemotherapy approach to cancer treatment, requiring the development of personalized cancer medicine strategy. Recently, high-throughput sequencing technology provide a vast of whole genome sequencing data for us, facilitating the research of combinatorial drug therapies with computational methods. Therefore, we utilized a model for precision drug combination prediction called PDCP using cancer patients’whole genome sequencing data and established a web server for application. Firstly, we mapped the somatic mutations and copy number variations in an individual’s genome into drug gene interactions. Secondly, druggable mutated genes were computed with a topological importance score in gene interaction network and genes with top topological scores was selected. Thirdly, the generated gene lists was combined pair-wised and gene pairs with high performance in the gene interaction network was generated. Finally, combinatorial drug regimens was assigned according to the generated gene pairs. We validated this approach with eight datasets provided by The Cancer Genome Atlas (TCGA). The results of the validation demonstrated that PDCP can correctly predict the drug combinations for individuals. |
PDF Full Text Request