Fig. 10From: Architecting the metabolic reprogramming survival risk framework in LUAD through single-cell landscape analysis: three-stage ensemble learning with genetic algorithm optimizationDissecting the malignant cells with high 3Â S-MMR score. (A) The development trajectory of malignant cells inferred by Monocle2. Malignant cells with high 3Â S-MMR scores most located in the roots of differentiation, and the malignant cells with low 3Â S-MMR scores mainly located in the middle and end-point state. (B) Heatmap of the 3Â S-MMR score-related genes in malignant cells along the pseudo-time. (C) Heatmap showing the different TFs activation between high and low 3Â S-MMR score malignant cells. (D, E) Top activities of TFs between high (D) and low 3Â S-MMR (E) score of malignant cells. RSS indicates Regulon Specificity Score. (F, G) Cellchat analysis of all cell types. Both interaction numbers and interaction strengths were showed. (H, I) Hierarchical plot showing the inferred intercellular communication network for SPP1 signaling pathway. (J) HE staining showing histologically distinct regions of stRNA samples. yellow: cancer region. (K) The spatial plot of 3Â S-MMR score intensity. (L) The distribution of different cell types in the spatial map was identified by the algorithm of RCTD.Back to article page