Fig. 4

Gene sets for candidate ILC- and IDC-specific TFs display coherent functional annotations and consistent expression changes in tumors. A Targets of EGR1, TEAD4, SOX2, and RUNX3_BCL11A, ILC-specific candidate TFs, showed significant upregulation in ILC tumors relative to IDC tumors (p value < 1e − 3, one-sided Kolmogorov–Smirnov test) compared to background genes. The upper panel depicts the upregulation of TF target gene expression in TCGA RNA-seq data, and the bottom panel depicts METABRIC microarray data. B Targets of FOXA1, HSF4, PBX3, and PITX1, IDC-specific candidate TFs showed significant upregulation of expression in TCGA and METABRIC data. The background genes were all genes identified in the gene expression dataset after removing low or non-expression genes. The yellow lines are empirical cumulative distribution functions (eCDF) for the target gene log2 fold changes between ILCs and IDCs. The blue lines are CDFs for background gene log2 fold changes between ILCs and IDCs. The p values are from the one-sided Kolmogorov–Smirnov (K–S) test between the target and the background distributions