Fig. 4. Regression analysis for transcriptomic signatures to predict clinical risk.
a Overview of processing and elastic net regression analysis of transcriptomic data in combination with FAERS-derived clinical risk scores. b Transcriptomic signature genes selected to predict cardiotoxicity risk score indicating their variable importance. c Observed and predicted risk scores from the elastic net cross- validation analysis (mean and standard deviation). d External validation of the signature for six kinase inhibitors: regorafenib (REG), sunitinib (SUN), ibrutinib (IBR), lenvatinib (LEN), nintendinib (NIN), and ceritinib (CER).