Figure 1.

Single-cell composition and expression characteristics of the pancreatic cancer liver metastasis microenvironment. (A) Distribution of cells in the pancreatic cancer liver metastasis dataset after UMAP dimensionality reduction and cell cycle correction. PT: Pancreatic cancer primary lesion tissue, LM: Pancreatic cancer liver metastasis tissue. (B) Seurat matrix showing the 18 cell subtypes (Cluster 0-17) in liver metastasis tissue. (C) Classification of 18 cell subtypes into 5 cell types through a combination of machine learning algorithms and manual annotation: Epithelial cells (Cluster 0, 1, 3, 6, 7, 10, 11, 12, 14, 15, 16), Macrophages (Cluster 2), Fibroblasts (Cluster 4, 5, 13, 17), T cells (Cluster 8), Endothelial cells (Cluster 9). (D) Classification of the pancreatic cancer liver metastasis microenvironment into three main components: stroma (fibroblasts, endothelial cells), epithelium (epithelial cells), and immune (T cells, macrophages) based on cell types. (E) Gene expression levels of individual cells, with darker colors indicating higher gene expression. (F) Copykat algorithm inference results for malignant cells, with red cell subtypes defined as malignant cells. (G-L) Proportions of various cell subtypes and cell types in different groups. (G) Proportions of 18 cell subtypes in 15 samples. (H) Proportions of 18 cell subtypes in pancreatic cancer primary lesion and liver metastasis. (I) Proportions of 5 annotated cell types in pancreatic cancer primary lesion and liver metastasis. (J) Proportions of the three main TME components in pancreatic cancer primary lesion and liver metastasis. (K) Proportions of the three main TME components in 15 samples. (L) Proportions of cells identified as malignant by the copykat algorithm in 15 samples.