Extended Data Fig. 3: Bulk RNA-seq analysis of transcriptional differences across different intestinal tissues and single-cell immune profiling of CD4⁺ and CD8⁺ T cells.

a, Dendrogram showing unsupervised hierarchical clustering of tissue samples from CRC patients and healthy donors that were run on the expression of immune-related genes. b, Kaplan Meier curves displaying differences of DFS between the two clusters in TCGA-COAD cohort. c, Scatter plots showing immune genes expression fold change (log2) between C1 and C2 (y-axis) against the corresponding values of ileum and colon (x-axis). Dots in red circle represent top immune genes with significant expression change on both axes. Pathways enrichment of the selected immune genes and the most enriched pathways are labeled on the top. d, Box plots showing normalized T cell activation pathway enrichment score in different clusters across CRC tumors and normal tissues, ordered by decreasing signature score. The number of samples in each group is listed below the boxes. e, Scatter plot showing the correlation of differentially expressed immune related genes between ileum and colon in CRC patients (y-axis) against the corresponding values in healthy donors (x-axis). f, The same UMAP as in Fig. 2a but with cells colored by TCR productivity and gene expression of TYROBP, LEF1, CD160. g, Bubble plot showing the expression of indicated genes for CD8⁺ T cells and IEL-T subtypes as defined in Fig. 2a. Dot size indicates frequencies of expressing cells, colored according to average expression levels. h, UMAP plot showing subtypes of CD4⁺ T cells. i, Bubble plot showing the expression of select marker genes for CD4⁺ T cells subtype as defined in (h). Dot size indicates frequencies of expressing cells, colored according to average expression levels. j, UMAP plot showing subtypes of CD8⁺ T cells from PBMC and LN. k, Bubble plot showing the expression of select marker genes for CD8⁺ T cells subtype from PBMC and LN. Dot size indicates frequencies of expressing cells, colored according to average expression levels. l, Box plots showing the paired comparisons of cell proportions of T_Reg cells, T_Naive cells, T_CM cells, T_FH cells, T_CTL cells and T_EM cells in CD4⁺ T cells among matched tissues from the same CRC patients (n = 10) and healthy donors (n = 3). m, Box plots showing the paired comparisons of cell proportions of GZMK⁺ eff. cells, T_Mem cells, T_Naive cells and MAIT cells in CD8⁺ T cells among matched tissues from the same CRC patients (n = 10) and healthy donors (n = 3). n, Box plots showing neoantigen reactivity for CD4⁺ (left) and CD8⁺ (right) cells across the indicated cell subtypes. o, Box plots showing CD4⁺ neoantigen reactivity of T_Reg and T_FH cells across the indicated tissue types in CRC patients and healthy donors. P values are determined by paired two-sided Mann-Whitney test (d, l, m and o). Box plots: center line, median; box limits, first and third quartile; whiskers, 1.5× interquartile range.