Extended Data Fig. 5: Data robusticity evaluation, tissue similarity and characterization of fibroblast subsets in CRC.
(a) Box plots (top and bottom quartiles with horizontal lines at the median) showing cell proportion in relative to all non-epithelial/malignant cells in different tissues after randomly excluding (Excl.) datasets. Each dot indicates a sample. From left to right, n = 28, 24, 24, 28, 28, 28, 28, 28, 28 (healthy), 10, 6, 6, 10, 10, 10, 10, 10, 10 (uninflamed), 15, 11, 11, 15, 15, 13, 15, 15, 15 (inflamed), 83, 83, 78, 68, 77, 83, 76, 71, 42 (paracancerous), 16, 16, 16, 16, 16, 16, 16, 16, 16 (polyp), 180, 180, 175, 154, 164, 180, 162, 163, 63 (tumor). (b) Box and violin plots (top and bottom quartiles with horizontal lines at the median) showing Bhattacharyya distance difference between tumors and polyps in different cell types, n = 100 for each cell type. (c) Lollipop plot showing upregulated pathways in monocyte/macrophages in tumors compared to polyp tissues. Numbers in the dots indicate gene counts matched to corresponding biological pathways. Over-representation analysis (BH adjustment). (d) UMAP plots showing fibroblast subsets composition (top) and distribution in tissues (bottom). n = 45581 cells. (e) UMAP plots showing expression patterns of POSTN (top) and CXCL12 (bottom). n = 45581 cells. (f) Lollipop plots showing upregulated pathways in CAF subsets. Numbers in the dots indicate gene counts matched to corresponding biological pathways. Over-representation analysis (BH adjustment).
