Research Paper Volume 15, Issue 14 pp 7146—7160
Combining WGCNA and machine learning to construct immune-related EMT patterns to predict HCC prognosis and immune microenvironment
Figure 2. Machine learning construction iEMT_score and prognostic value verification. (A) SVM-REF algorithm to screen the optimal feature subset. (B) Lasso algorithm adjusts feature selection and constructs iEMT_score. (C) Kaplan-Meier analysis of iEMT_score subgroups in the TCGA-HCC cohort (training cohort). (D) Kaplan-Meier analysis of iEMT_score subgroups in the ICGC-Japan-HCC cohort (validation cohort). Kaplan-Meier survival curve analysis of iEMT_score in different ages (E, F), genders (G, H), tumor grades (I, J) and stages (K, L).