Construction of a prognostic model based on glycosylation-related genes in colorectal cancer. (A)Forest plot showing the association between four glycosylation-related genes and overall survival in colorectal cancer patients, including hazard ratios with 95% confidence intervals and p-values. (B) Tuning parameter (λ) selection in the LASSO regression model by minimum criteria. The partial likelihood deviance curve was plotted versus log2(λ), with the optimal λ value of 0.0015 determined by 10-fold cross-validation. (C) LASSO coefficient profiles of the four prognostic glycosylation-related genes at the optimal λ value. (D) Distribution of GRGS risk scores, survival status, and expression patterns of the four signature genes in the TCGA colorectal cancer cohort. (E and F) Boxplots comparing GRGS risk scores between tumor tissues and normal tissues in the TCGA (E) and GSE39582 (F) cohorts. (G) Kaplan-Meier curves for overall survival comparing the high-risk and low-risk groups stratified by the GRGS model. (H and I) Kaplan-Meier curves for overall survival in (H) GSE17536 and (I) GSE39582 datasets comparing patients in GRGS-High and GRGS-Low risk groups. (J) Kaplan-Meier curves for disease-specific survival in GSE17536 dataset comparing patients in GRGS-High and GRGS-Low risk groups. (K) Kaplan-Meier curves for relapse-free survival in GSE39582 dataset comparing patients in GRGS-High and GRGS-Low risk groups. (L) Kaplan-Meier curves for overall survival in GSE87211 datasets comparing patients in GRGS-High and GRGS-Low risk groups. ***, p < 0.001.
|
