Fig. 7: Dissection of tumor microenvironment based on the PCDI signature.
From: Machine learning reveals diverse cell death patterns in lung adenocarcinoma prognosis and therapy
a UMAP plots for 80059 cells in two LUAD single-cell datasets. b Heatmap showing the top differentially expressed genes (Wilcoxon test) in each cell type. c Feature plots present the typical marker gene expression for each cell type. d The distribution and expressions of ten model genes across different cell types. e Heatmap plots showing the CNV signals for each epithelial cluster. f Box plots of the score for each epithelial cluster. g U-MAP plot showing the three subclusters of epithelial cells. h The trajectory of the differentiation process is colored by state and cell type. i The correlation between CNVscore and CDIscore. j Comparison of the CDIscore between low and middle/high malignant subclusters.
