Fig. 1

Global unsupervised clustering of 1148 breast samples using RNA-Seq data. Applying the K-Means algorithm using K = 5 on the RNA-Seq dataset yielded a partition exhibiting moderate agreement with PAM50 labels and the three immunohistochemical markers. Notably, luminal-A samples were split between a rather homogenous cluster 2 and cluster 1, which is composed of a mix of luminal-A and luminal-B. a K-Means clusters. b PAM50 calls. c Estrogen receptor (ER) status. d Progesterone receptor (PR) status. e Human epidermal growth factor receptor 2 (HER2) status