Figure 2.

Visual analysis of the data sets. A_old is plotted in green, A_new is plotted in orange, the first row (A–C) shows hERG data, the second row (D–F) shows Na_V, and the third row (G–I) shows the hERG augmented data set. Panels A and D plot the truncated SVD of the signature descriptors, explaining 13.28% and 13.62% of the total variance in data. Panels B and E plot the computed signature descriptors using t-SNE dimensionality reduction. Panels C and F plot the distribution of measured assay values, where the dashed lines show the mean for each assay. Panels G and H display the augmentation made to the hERG measurements in A_old, expressed in IC₅₀; the dashed line in panel G corresponds to a 1:1 relation between the original and augmented data (i.e., no change). The dashed line in panel H is the mean value of the plotted distribution.