Fig. 3

Construction of prognostic models. (A) The forest plot of univariate Cox analysis. (B) LASSO regression analysis for biomarkers. (C) PH assumption test. (D) Risk curve and heat map of gene expressions in high- and low-risk cohorts of the training set. (E) Kaplan–Meier curves for the high- and low-risk cohorts in the training set. (F) ROC curve in the training set. (G-H) The expression levels of ATP6V0D1 and HDAC6 in GSE16088 (G) and GSE14359 (H). LASSO, least absolute shrinkage and selection operator; PH, proportional hazards; ROC, receiver operating characteristic; HR, hazard ratio; lower 95%CI and upper 95%CI represent the 95% confidence intervals of the risk values