Fig. 1

Unsupervised hierarchical clustering as well as t-distributed stochastic neighbor embedding (t-SNE) to compare DNA methylation signatures across different breast cancer subtypes from The Cancer Genome Cohort (n = 422) as well as our own analyses (n = 36). a Unsupervised hierarchical clustering reveals four distinct clusters, representing normal breast tissue (“Normal Cluster”), tumors with hormone receptor expression (“HR+ Cluster 1” and “HR+ Cluster 2”) as well as TNBCs (“TNBC Cluster”). Molecular subtype classification based on results from immunohistochemistry (IHC) and/or HER2 in situ hybridization as well as RNA sequencing using the AIMS classification are shown below the heatmap. b General annotation of normal breast tissue and different breast cancer subtypes from IHC and/or in situ hybridization in a t-SNE plot. c and d t-SNE plot showing the distribution of estrogen (c) and progesterone (d) receptor expression. e t-SNE visualizing the HER2 amplification status. f Correlation between t-SNE analysis and hierarchical clustering. Abbreviations: ER estrogen receptor, PR progesterone receptor