. 2004 Feb;13(2):400–411. doi: 10.1110/ps.03348304

Table 7.

The success rate and the average Z-score of different potentials using a subset of the Levitt’s multiple decoy sets

Method^a	TE-13	MJ	GKS	BT	HL	BJ
# Correct/Total^b	13/15	11/25	9/25	9/25	8/25	15/25
〈Z score〉	3.53 ± 1.14	2.82 ± 2.27	2.36 ± 2.53	2.65 ± 2.37	2.67 ± 2.02	2.75 ± 2.10
Method	XCJ	Errat	ProsaII	Verify3D	DFIRE-SCM
Correct/Total	11/19	11/25	15/25	10/25	19/25
〈Z score〉	2.72 ± 1.82	4.04 ± 2.45	3.05 ± 1.63	2.40 ± 1.74	4.52 ± 1.75

^a Energy functions listed: Tobi and Elber (TE-13) (Tobi and Elber 2000), Miyazawa and Jernigan (MJ) (Miyazawa and Jernigan 1999), Godzik, Koliniski, and Skolnick (GKS) (Godzik et al. 1995), Betancourt and Thirumalai (BT) (Betancourt and Thirumalai 1999), Hinds and Levitt (HL) (Hinds and Levitt 1992), Bahar and Jernigan (BJ) (Bahar and Jernigan 1997), Xiang, Chang and Jie (XCJ) (Li et al. 2003), Colovos and Yeates (Errat) (Colovos and Yeates 1993), Sippl(ProsaII) (Sippl, 1993), Eisenberg, Luthy, and Bowie (VERIFY-3D) (Eisenberg et al. 1997), and DFIRE-SCM (this work).

^b The number of correctly ranked as number one in the total of 25 multiple decoys used in Tobi and Elber (2000). The decoy sets include 4state_reduced (1ctf,1r69,1sn3,2cro,4pti,4rxn), fisa (1fc2,1hdd-C,2cro), fisa_casp3 (1bg8-A,1bl0,1jwe), lattice_ssfit (1etf,1dkt-A, 1fca,1nkl,1pgb,1trl-A) and lmds (1ctf,1dtk,1fc2-C,ligd,1shf-A,2cro,2ovo).