TB-IECS: an accurate machine learning-based scoring function for virtual screening

Table 6 Model performance on DUDE

Targets	Chemplp_smina_nn_smo			Smina_nn (TB-IEC Score)			Glide_SP		DUDE_Dock
Targets	ROAUC	F1_Score	EF_1%	ROAUC	F1_Score	EF_1%	ROAUC	EF_1%	ROAUC	EF_1%
akt1	0.976	0.733	50.896	0.993	0.833	50.610	0.584	5.054	0.720	29.4
ampc	0.892	0.462	36.228	0.979	0.519	41.610	0.558	1.057	0.789	8.3
cp3a4	0.941	0.618	24.970	0.993	0.789	25.000	0.597	2.985	0.631	2.4
cxcr4	0.994	0.927	11.063	1.000	0.986	11.090	0.639	2.885	0.902	17.5
gcr	0.997	0.727	50.286	0.995	0.742	49.060	0.746	18.074	0.439	8.9
hivpr	0.995	0.903	17.894	0.999	0.970	17.900	0.531	2.359	0.596	4.7
hivrt	0.944	0.662	27.855	0.959	0.746	27.900	0.563	7.125	0.644	6.5
kif11	0.955	0.660	41.273	0.973	0.825	40.750	0.635	16.524	0.769	34.5

The best performance under the perspective of three metrics for each target are shown in bold

ISSN: 1758-2946