Fig 1.
Extraction and WGCNA of differentially expressed hypoxia-related genes (HGs).
(A) The multiple GSEA for 10 hypoxia-related gene sets are based on the TCGA datasets. (B) The volcano plot for the 199 HGs in the TCGA datasets. (C) The volcano plot for the 168 HGs in the GSE14520 datasets. (D) The identification of the 83 overlapping HGs ground on GSE14520 and TCGA. (E) Sample dendrogram and trait heatmap for the detection of outliers ground on the TCGA datasets. (F) Scale-free topology model fit (left) and mean connectivity (right) for the appropriate soft threshold power. The power selected was 12. (G) Clustering dendrograms of the co-expression network modules of HGs. (H) Module‑trait relationships between the tumor traits and module, correlation coefficient, and p-values are shown.
Fig 2.
Construction and validation of prognostic risk signature of HGs for HCC patients.
(A) Forest plot of univariate Cox regression analyses for 25 HGs related to overall survival ground on the TCGA dataset. (B) The partial likelihood deviance plot. (C) The Lasso regression coefficient profiles. (D) Multivariate Cox regression analyses of 3 core HGs. (E) Cox coefficients distribution of the three core HGs.(F-H) Survival status of patients, risk plot distribution, and heatmap of expression of 3 core HGs in the (F) training, (G) testing, and (H) GSE14520 cohorts.
Fig 3.
Evaluation of predictive power of the selected 3 HGs signature.
(A-C) Kaplan–Meier survival curves for the risk signature based on the training, testing and GSE14520 cohort. (A) 186 cases in the training cohort. (B) 184 cases in the testing cohort. (C) 221 cases in the GSE14520 cohort. (D-F) Receiver operating characteristic (ROC) curves for the risk signature in the training, testing and GSE14520 cohort. (D) training cohort. (E) testing cohort. (F) GSE14520 cohort.
Fig 4.
Identification of the risk signature with the overall survival and clinicopathological characteristics of HCC patients ground on the TCGA dataset.
(A) Relationships between the risk score and clinicopathological characteristics of HCC patients. (B) Forest maps of the univariate and multivariate Cox regression analysis between the risk score and clinical characteristics. (C) Nomogram predicting the survival rate at 1, 3, 5 years for HCC patients, *p < 0.05, **p < 0.01, and ***p < 0.001. (D) Calibration plots for the nomogram.
Fig 5.
The prognostic ability of the 3 HGs signature for overall survival in multiple HCC subtypes.
Kaplan–Meier curves for OS prediction in HCC subtypes of (A) age < = 55, (B) age >55, (C) Men, (D) Women, (E) Stages I–II, (F) Stages III–IV in the TCGA cohort, (G) age < = 55, (H) age >55, (I) Men, (J) Women, (K) Stages I–II, (L) Stages III in the GSE14520 cohort.
Fig 6.
Association of tumor immune cell infiltration with the risk score in the TCGA cohort.
(A) The bar graph of relative proportions of the 22 immune cells between the high and low-risk subgroups. (B) The bar graph of difference in composition of the 22 types of immune cells between two risk subgroups, *p < 0.05, **p < 0.01, and ***p < 0.001. (C) The correlation between the risk score and naive B cells, resting memory CD4 T cells, monocytes, and M0 macrophages. (D) The bar graph of the difference in the enrichment scores of 16 types of immune cells between two risk subgroups (E) The multiple GSEA for significant immune pathways based on the TCGA dataset.
Fig 7.
Characteristics of risk score with tumor somatic mutation and TMB in the TCGA dataset.
(A) The waterfall plot of tumor somatic mutation in the high-risk subgroups(B)and the low-risk subgroups. (C) TMB distribution in the high and low-risk groups. (D) Kaplan–Meier survival curves for the HCC patients between high-TMB high-risk, high-TMB low-risk, low-TMB high-risk, and low-TMB low-risk subgroups.
Fig 8.
Association of risk score with chemotherapy sensitivity and the expression of immune checkpoints in the TCGA cohort.
(A-D) Estimated IC50 for (A) LY317615, (B) PF−562271, (C) Pyrimethamine, and (D) Sunitinib in high and low-risk subgroups. (E) The expression level of possible immune checkpoints in high and low-risk groups.