UnCorrupt SMILES: a novel approach to de novo design

Table 2 Properties of target-directed RNN and SMILES exploration outputs

SNN is the similarity to the nearest neighbor of 10,000 molecules from the generated sets compared to the set of 1627 known AURKA and AURKB ligands. Fragment and scaffold similarity are calculated by comparing the frequency distribution of different fragments or scaffolds compared to the known ligands. KL divergence describes the similarity of the physicochemical property distributions of 10,000 molecules from the generated sets compared to the know ligands

ISSN: 1758-2946