Fig. 2. Interactions of T-cell dysfunction signature genes with immune checkpoints in uterine corpus endometrial carcinoma (UCEC).

ImogiMap-based assessment of combinatorial interactions for T-cell dysfunction signature genes in UCEC, and therapeutically actionable ICPs, based on their associations with the IFNG gene expression. The analysis is based on RNA expression data from tumors of 370 patients. a Combined action score robustness. A normalized root mean square deviation (RMSD) is calculated for each combined action score through random sampling (1000x) of sub-cohorts with 70% coverage of the complete cohort. Scores with high robustness, $-\log \left(R\right)>0$ (red horizontal line) are selected for further analysis (see “Methods”). b Statistical significance of combined action scores. An FDR (Benjamini-Hochberg (BH) method) corrected $Q$-value based on the Wilcoxon signed-rank test is calculated for robust scores. Scores with $Q_{\_\mathrm{Significance}}<0.1$ are selected for further analysis. c A BH-corrected Q-value for specificity is calculated for each robust and significant interaction (see “Methods”). Scores with $Q_{\_\mathrm{Specificity}}<0.1$ are selected for further analysis. d The graphical network representing robust, significant, and specific combinatorial associations with IFNγ levels (represented by IFNG gene expression). Red (Blue) edges represent upregulation (downregulation) of IFNG. Dark red (Dark blue) vertices identify overexpression (low-expression) of ICP genes and orange (blue) vertices identify overexpression (low-expression) of T-cell dysfunction signature genes. e Levels of IFNG gene expression in TCGA UCEC samples, stratified based on SERPINB9 and CTLA4 levels. p values from Wilcoxon signed-rank test indicate the statistical significance of differences in IFNγ levels within the stratified sub-cohorts. f Levels of IFNG gene expression in TCGA UCEC samples, stratified based on CD86 and CD70 levels. g Kaplan–Meier survival curve for UCEC patients stratified by low/high expression of CD70 and CD86.