Fig. 2

Identification of an at-risk CRC subpopulation stratified by peroxisome score and TIM3 expression. a Pairwise correlations of the peroxisome score versus expression of individual immune checkpoint genes in TCGA cohort. Pearson’s rho (r) and statistical difference (p) are shown in each graph. b, c Scatter plots of peroxisome score and log2-transformed TIM3 gene expression values are shown for TCGA dataset. MSI (blue diamond) and MSS (black circle) status are labeled for CRC patients. The median value of TIM3 expression is indicated in a gray dashed line. The log-rank test score patterns and candidate cut-offs (blue and red arrows) for high peroxisome and low peroxisome groups are shown in panels b and c, respectively. d, e Kaplan–Meier curves show the OS for the optimal cut-off of TIM3 expression in TCGA CRC subgroups with high peroxisome score and low peroxisome score, respectively. Using the median value of peroxisome score and these two candidate cut-offs, we separated patients into four different groups labeled with their population fraction in percentage (group I-IV). Kaplan–Meier curves are plotted for the risk groups in high peroxisome and low peroxisome groups, respectively. P-value was calculated by the log-rank test and shown for each plot