Seven Hub Genes Predict the Prognosis of Hepatocellular Carcinoma and the Corresponding Competitive Endogenous RNA Network

Figure 3

Identification of hub genes by least absolute shrinkage and selection operator (LASSO) and univariate Cox models. (a) The top 14 genes are shown with AUC>0.80 in TCGA and GSE76427. Genes that have been upregulated are shown in red, whereas genes that have been downregulated are shown in blue. (b) The correlation between the number of trees and the error rate. (c) A random forest model was constructed to determine the importance of the 14 genes in an order. (d) LASSO coefficient profiles of the 7 feature genes. (e) 10-fold cross-validation of parameter selection in the LASSO analysis. (f) Univariate analysis of feature genes. The feature genes with hazard ratios >1 are poor prognosis genes, and hazard ratios <1 are protective genes. TCGA, The Cancer Genome Atlas; AUC, area under the curve.

(a)

(b)

(c)

(d)

(e)

(f)