Multi-objective de novo drug design with conditional graph generative model

Table 3 Performance of graph based and SMILES based model on scaffold diversification tasks

Condition (\({\mathbf{c}}\))	\(R_0\)	\(I_0\)	Model	% valid	\(R_{\mathbf{c}}\)	\(EOR_{\mathbf{c}}\)	Diversity (\(I_{\mathbf{c}}\))
scaffold 1	\(7.9\times 10^{-5}\)	0.46	Graph	0.931 ± 0.008	0.86 ± 0.03	10865	0.45 ± 0.01
scaffold 1	\(7.9\times 10^{-5}\)	0.46	SMILES	0.924 ± 0.005	0.87 ± 0.01	10976	0.46 ± 0.01
scaffold 2	\(1.1\times 10^{-4}\)	0.50	Graph	0.900 ± 0.016	0.77 ± 0.04	6972	0.47 ± 0.02*
scaffold 2	\(1.1\times 10^{-4}\)	0.50	SMILES	0.896 ± 0:011	0.84 ± 0.01*	7607	0.44 ± 0.01
scaffold 3	\(7.9\times 10^{-5}\)	0.56	Graph	0.940 ± 0.019*	0.56 ± 0.08**	7086	0.60 ± 0.02
scaffold 3	\(7.9\times 10^{-5}\)	0.56	SMILES	0.898 ± 0.024	0.37 ± 0.07	4623	0.59 ± 0.03
scaffold 4	\(5.8\times 10^{-3}\)	0.82	Graph	0.982 ± 0.001***	0.88 ± 0.01	151	0.815 ± 0.001
scaffold 4	\(5.8\times 10^{-3}\)	0.82	SMILES	0.969 ± 0.002	0.88 ± 0.00	151	0.819 ± 0.00***

Results are reported as \(Mean \pm SD\). The best performance in each metric is highlighted in italics face. Paired t-tests are carried out for the difference between the graph and SMILES based method (*** for \(p\le 0.001\), ** for \(p\le 0.01\) and * for \(p\le 0.05\))

ISSN: 1758-2946