Figure 5

Model constructed using the METABRIC dataset and its association with clinical and genetic variables. (A) Breast cancer progression tree. Each node represents a cluster and the node size is proportional to the number of samples in the corresponding cluster. Nodes are color-coded based on the PAM50 labels of the majority of the samples in the node. (B,C) Molecular grade and CNA frequency were highly correlated with the N-B (first column) and N-H (second column) progression branches. (D) Spearman’s rank correlation analysis of histological grade, molecular grade, mutation rate and patient age with the two main progression paths. The numbers in parenthesis are P values. CNA, copy number alteration; N-B, normal through luminal to basal phenotype; N-H, normal through luminal to HER2+ phenotype; TCGA, The Cancer Genome Atlas.