Fig. 4

Construction and validation of the FRG-based prognostic signature. A LASSO regression analysis to select the optimal penalty lambda (λ). B LASSO coefficients of FRGs corresponding to log(λ). C Alluvial diagram revealing the distribution of patients in the two FRG subtypes and two risk groups. Scatter plots of the risk score, and survival status and time, and heatmap for expression of two genes in the training set (D), testing set (E), and entire (F) set. K–M curve of high- and low-risk groups in the training set (G), testing set (H), and entire set (I). ROC curves of the risk score and clinical characteristics in the training set (J), testing set (K), and entire set (L)