Fig. 1: Distribution of breast cancer molecular subtypes defined by topological data analysis (TDA) signatures across ancestries.

a Heatmap of expression of PAM50 genes organized by TDA signature classes in TCGA breast cancer and RA-QA cohort. Samples are annotated by TDA signature class (upper annotation bar) and classical PAM50 intrinsic molecular subtype (lower annotation bar). The combination patterns of upregulated expression of five distinct gene groups defining each TDA class are summarized in a table on the right (Summary TDA). b Reclassification of breast cancer samples from classical PAM50 intrinsic molecular subtypes (upper part of circos) to TDA signature classes (lower part of circos) in TCGA and RA-QA breast cancer cohorts. c Stacked bar chart of distribution of TDA classes by ancestry. d Kaplan–Meier plots showing overall survival (upper panels) and disease-specific survival (lower panels) by ancestry. Difference between the survival of patients with European and African ancestry is shown for the complete TCGA breast cancer cohort (left), patients with TNBC according to hormone receptor status (middle left), patients with PAM50-defined basal breast cancer (middle right), and patients with tumors classified as BasalMyo by TDA classification (right). Censor points are indicated by vertical lines.