Fig. 8.

t-distributed stochastic neighbor embedding (t-SNE) of: (a) the sets of 100 compounds selected with the different selection methods, (b) sets of 500 ligands of different pharmaceutical targets, and (c) sets of cytochrome P450 2C9 ligands and inactive compounds from PubChem Bioassay 1851. The table summarizes the number of compounds in the sets, as well as computation times and average similarities (averaged over the 19 non-weighted similarity metrics, and, for the binary comparisons, also over all possible compound pairs)