EDock: blind protein–ligand docking by replica-exchange monte carlo simulation

Table 2 Summary of docking results on the targets with binding site error < 8 Å

Datasets	Methods	RMSD (Å)		Center distance (Å)		Ave RMSD < 2.0 Å
Datasets	Methods	Ave	Med	Ave	Med	Ave RMSD < 2.0 Å
DUDE (76)	EDock	1.45	0.33	0.88	0.23	64
	DOCK6	2.03	0.54	1.17	0.37	57
	Vina	3.92	2.73	2.05	1.13	36
	Vina_blind	5.15	1.15	3.50	0.92	40
COACH (315)	EDock	2.17	0.43	1.21	0.33	221
	DOCK6	2.84	0.90	1.56	0.51	190
	Vina	3.92	2.96	1.92	1.10	139
	Vina_blind	7.37	6.07	5.65	2.83	101

Since 4 (and 8) test targets failed in Vina (and DOCK6), only 391 targets (76 from DUDE and 315 from COACH), on which all programs have a result, are shown. “Ave” and “Med” represent the average and median values respectively. “EDock”, “DOCK6” and “Vina” represent the results starting with the same binding sites predicted by COACH, while “Vina_blind” refers to the blind docking result of AutoDock Vina with searching the whole receptor space. The best performance is highlighted in italic font in each category

ISSN: 1758-2946