Table 2.

Performance comparison between our model and existing methods on the DUD-E dataset

Model	AUC	0.5% RE	1.0% RE	2.0% RE	5.0% RE
NNscorea	0.584	4.166	2.980	2.460	1.891
RF-scorea	0.622	5.628	4.274	3.499	2.678
Vinaa	0.716	9.139	7.321	5.811	4.444
3D-CNNa	0.868	42.559	26.655	19.363	10.710
PocketGCNa	0.886	44.406	29.748	19.408	10.735
DrugVQAa	0.972 ± 0.003	88.17 ± 4.88	58.71 ± 2.74	35.06 ± 1.91	17.39 ± 0.94
MINN-DTI	0.992 ± 0.007	175.89 ± 12.02	90.77 ± 5.81	46.49 ± 2.63	19.10 ± 0.71

Note: The first row, the percentage before RE is the given threshold of FPR. Best results of the corresponding experiments were represented in bold.

^aMeans results obtained from the article (Zheng et al., 2020).