Research Paper Volume 16, Issue 2 pp 1581—1604
Identification of a basement membrane gene signature for predicting prognosis and estimating the tumor immune microenvironment in prostate cancer
Figure 4. BMRM construction and prognostic analysis. (A) Selection of the optimal penalty parameter (λ) in the LASSO Cox regression model. (B) Gene coefficient spectrum of the 16 RFS-associated BMGs in the TCGA-PRAD cohort in the LASSO Cox regression analysis. (C) Forest plot of multivariate analysis showing 5 genes (THBS2, MMP26, COL4A6, MMP21 and SDC1) as independent prognostic factors for RFS in PCa patients. (D, E) BMRS plots from the TCGA training set. (F) Heat map of the two BMRS groups and correlation analysis of clinicopathological characteristics. (G) Kaplan-Meier analysis of RFS for PCa patients in both risk groups. (H) AUC shows the accuracy of BMRM in predicting 1-, 3-, and 5-year RFS in the TCGA training set.