Fig. 4

Gene expression and cell type abundance in an independent cohort. a Distribution of HER2-high breast tumors (n = 53) vs HER2-low (n = 317) in DCIS and IDC. HER2-high samples were defined as having log2(HER2-expression) > 11. The pink area indicates proportion of HER2-high tumors, and the gray area indicates HER2-low. Chi square test indicated significantly different distribution of HER2-high tumors between DCIS and IDC (X2 (df = 1) = 25,71, p < 0.001). b Comparison of differentially expressed genes in HER2-high and HER2-low groups. Numbers indicate significantly differentially expressed genes in each group (FDR < 0.05). Selected genes are shown in corresponding boxes. c Immune cell deconvolution in HER2-high DCIS and invasive tumors. Each bar represents one sample, and the height of the colored bars represent the relative estimated abundance of different immune cells. d Top enriched Hallmark gene sets correlated with B-cell abundance in HER2-high DCIS. Normalized Enrichment Scores for the top five signatures positively correlated with B-cell abundance (NES > 0) and the top five signatures negatively correlated with B-cell abundance (NES < 0) in HER2-high DCIS. Colored bars indicate significant signatures (FDR < 0.1)