Contextualization of drug-gene relationships through integration of public knowledgebases
In the study of therapeutic drugs, genes and drugs are often theorized to have some relationship based on prior analysis. However, it is currently difficult to find how these drugs and genes might be related based on current literature and research. In this project, we remedy this problem by constructing context networks built with protein and drug interactions from the STITCH and STRING interaction databases. STITCH and STRING are sister databases that store drug-protein interactions and protein-protein interactions, respectively. Using these two databases, we’ve created an interface that allows researchers to explore the proteins associated with both the gene and the drug, enabling the researcher to contextualize the drug-gene pair with current research and knowledge.
To demonstrate the utility of these context networks, we applied them to the results of the EDDY-CTRP study. The Cancer Therapeutics Response Portal (CTRP) provides drug-response measurements for more than 400 small molecules using cell lines characterized by the Cancer Cell Line Encyclopedia. Using the CTRP dataset and EDDY, an algorithm that detects differential genetic statistical dependencies, we identified potential drug-mediator pairs. Mediators were identified as genes that showed significant change in genetic dependencies between the drug sensitive and drug non-sensitive cell lines and were predicted to be good candidates for understanding small molecule mechanisms and improving therapeutic effect. Using context networks, we were able to improve the interpretability of these results by linking the drugs and mediators with genes associated with both the drug and the mediator. We hope that with these networks, other researchers will be able to use the EDDY-CTRP results to generate hypotheses on how the mediators impact drug sensitivity.