Fig. 1.

Maximum similarity of each molecule in the test sets to the molecules in the training set. The similarity was calculated using the Tanimoto coefficient similarity on the topological fingerprints of the molecules