Fig. 8

3 S-MMR score’s ability to predict immunotherapy efficacy. (A-F) Violin plot of TIDE (A), Dysfunction (B), Exclusion (C), CD8 (D) MDSC (E), and Merck18 (F) score. (G) The submap algorithm predict the response of high and low 3 S-MMR score groups to CTLA4 and PD-1 inhibitors. (H) Boxplot of relative expression levels at immune checkpoints profiles between the high and low 3 S-MMR score groups patients. (I-N) Differences in 3 S-MMR score between immunotherapy responders and non-responders in the GSE126044 (I-J), GSE35640 (K-L), and GSE78220 (M-N) cohorts. (O-P) T-SNE reduction maps the distribution of cells from SD and PR patients (O), and the distribution of 3 S-MMR score (P) in the GSE207422 dataset. (Q) Violin plot of 3 S-MMR score between SD and PR patients in the GSE207422 dataset. (R) Tissue preference of high and low 3 S-MMR groups estimate by R_o/e in the GSE207422 dataset. (S-T) T-SNE reduction maps the distribution of cells from SD and PR patients (S), and the distribution of 3 S-MMR score (T) in the GSE145281 dataset. (U) Violin plot of 3 S-MMR score between SD and PR patients in the GSE145281 dataset. (V) Tissue preference of high and low 3 S-MMR groups estimate by R_o/e in the GSE145281 dataset. Abbreviation: *P < 0.05; **P < 0.01; *** P < 0.001