Fig. 3
From: Knowledge-guided gene prioritization reveals new insights into the mechanisms of chemoresistance
The performance of drug sensitivity prediction based on ProGENI compared to the baseline methods using the LCL (a, b) and GDSC (c, d) datasets for a selection of drugs. The y-axis shows the SPCI corresponding to ProGENI while the x-axis shows the SPCI corresponding to the baseline. Each point in the scatter plot corresponds to one random choice of training/test set. The color of each point represents the density of points in that region: a dark red color on a point means that the point is surrounded by many other points, while a blue color on a point means that the point is isolated. The FDR is calculated using a one-sided Wilcoxon signed rank test corrected for multiple tests. a Performance of ProGENI-SVR versus PCC-SVR for the LCL dataset. b Performance of ProGENI-SVR versus EN-SVR for the LCL dataset. c Performance of ProGENI-SVR versus PCC-SVR for the GDSC dataset. d Performance of ProGENI-SVR versus EN-SVR for the GDSC dataset