Fig. 7

LR score for quantifying interaction. A. According to the results of the previous single-cell analysis. Schematic diagram showing the interaction network of ligands and receptors between MACRO-SPP1 (SPP1 + TAMs) and CD8T-C4 (exhausted CD8 + T cells). B. Pie chart showing ligand and receptor expression in tumor/normal tissues (TCGA-STAD) C. Bubble plot showing the differential expression of ligands and receptors in tumor/normal tissues (TCGA-STAD). D. Forest plot showing LR scores constructed based on the above ligands and receptors using LASSO and Cox regression methods (TCGA-STAD). E. Radar chart showing hazard ratio values based on univariate/multivariate Cox regression. The LR score had the highest multivariate and univariate HR value (TCGA-STAD). F–I. Kaplan‒Meier curve showing that the LR score is correlated with OS, PFS, DSS, and DFS (TCGA-STAD). J. The receiver operating characteristic (ROC) curve shows that the LR score has good evaluation performance (TCGA-STAD).