Fig. 2

Hierarchical clustering of INPs and reference compounds based on their median logGI50 values across NCI60 cell lines. The tree was inferred using the UPGMA (‘average’) method and was based on Euclidean distances. The tree is presented as an unrooted radial phylogram. The scale in the top left corner is provided for the branch length, which were derived from Euclidean distances. Clustered products are displayed with sparse labeling, in which only a random subset of INP labels is displayed. Detailed information about the INPs in each subtree is provided in Table 1 and Supplementary Table 4