Fig. 5
From: A risk model of gene signatures for predicting platinum response and survival in ovarian cancer
Construction and validation of the prognostic risk model. A. Cross-validation likelihood filters the lambda parameter (20.88803) when cvl takes the maximum value (− 771.2244). B Based on the L1-penalized regularized regression algorithm, the optimal prognostic gene coefficient distribution line for Cox-PH model screening. C Prognostic prediction by the prognostic risk model in the TCGA training dataset. D Prognostic prediction by the prognostic risk model in the GSE63885 validation set. E The correlation between the K-M curve of the platinum response status and the prognostic model prediction