Fig. 3
Comparison of ROC and LOCC in evaluating prognostic biomarkers A A ROC curve is plotted and the AUC was calculated for TCGA SARC E2F1. B LOCC was plotted and scored for SARC E2F1. C The most significant cutoff was selected and a Kaplan–Meier plot was graphed to illustrate the best stratification according to E2F1 in SARC. D A ROC curve is plotted and the AUC was calculated for TCGA PAAD E2F1. E LOCC was plotted and scored for PAAD E2F1. F The most significant cutoff was selected and a Kaplan–Meier plot was graphed to illustrate the best stratification according to E2F1 in PAAD. G A ROC curve is plotted and the AUC was calculated for TCGA KIRP E2F1. H LOCC was plotted and scored for KIRP E2F1. I The most significant cutoff was selected and a Kaplan–Meier plot was graphed to illustrate the best stratification according to E2F1 in KIRP. J A ROC curve is plotted and the AUC was calculated for TCGA LGG E2F1. K LOCC was plotted and scored for LGG E2F1. L The most significant cutoff was selected and a Kaplan–Meier plot was graphed to illustrate the best stratification according to E2F1 in LGG. Abbreviations and symbols: TCGA – The Cancer Genome Atlas, SARC – Sarcoma, PAAD – Pancreatic ductal adenocarcinoma, KIRP – Kidney renal papillary cell carcinoma, LGG – low grade glioma, pl is − log (p value) at most significant cutoff, Rs is percentage of cutoffs that have a highly significant p value (p < 0.01), HRp is the HR at the most significant cutoff