. Author manuscript; available in PMC: 2021 Aug 11.

Published in final edited form as: J Chem Theory Comput. 2020 Jul 6;16(8):5334–5347. doi: 10.1021/acs.jctc.0c00476

Table 3:

Summary of docking results measured by the number and the portion of targets in each set that have an acceptable near-native top-1, 3, 5, or 10 prediction.

	ZDOCK (Starting Point)				PSO				BAL

Dataset (size)	N₁ (F₁)	N₃ (F₃)	N₅ (F₅)	N₁₀ (F₁₀)	N₁ (F₁)	N₃ (F₃)	N₅ (F₅)	N₁₀ (F₁₀)	N₁ (F₁)	N₃ (F₃)	N₅ (F₅)	N₁₀ (F₁₀)
Training (50)	6 (12%)	11 (22%)	13 (26%)	17 (34%)	9 (18%)	15 (30%)	17 (34%)	20 (40%)	14 (28%)	19 (38%)	20 (40%)	22 (44%)
Test (126)	20 (16%)	29 (23%)	33 (26%)	41 (33%)	26 (21%)	33 (26%)	39 (31%)	45 (36%)	32 (25%)	40 (32%)	42 (33%)	50 (40%)
CAPRI (15)	2(13%)	2 (13%)	3 (20%)	4 (27%)	3 (20%)	3 (20%)	4 (27%)	5 (33%)	3 (20%)	4 (27%)	5 (33%)	5 (33%)