Skip to main content
. 2010 Jun;17(6):783–798. doi: 10.1089/cmb.2009.0235

Table 4.

Performance Comparison between Our CRFFolder and Skolnick's TOUCHSTONE-II

 
 
 
TOUCHSTONE
CRFFolder
Target Size Class BestCluster BestCluster Best 1% 2%
1bw6A 56 α 4.79 (2/3) 3.82 (3/3) 2.75 3.38 3.54
1lea 72 α 5.69 (5/5) 4.10 (5/7) 3.41 4.19 4.48
2af8 86 α 11.07 (5/6) 8.9 (12/19) 7.07 8.53 8.97
256bA 106 α 3.61 (2/3) 2.75 (6/11) 2.50 3.45 3.70
1sra 151 α 10.71 (3/12) 13.95 (17/25) 10.82 13.76 14.24
1gpt 47 α β 6.30 (1/25) 5.55 (42/67) 4.34 5.20 5.47
1kp6A 79 α β 10.01 (8/14) 7.99 (1/7) 6.29 7.51 7.81
1poh 85 α β 9.10 (5/9) 8.84 (5/10) 7.49 8.70 9.04
1npsA 88 α β 6.89 (33/34) 9.91 (41/57) 7.87 9.19 9.66
1t1dA 100 α β 8.96 (7/13) 9.22 (10/13) 6.51 9.51 9.94
1msi 66 β 7.72 (19/28) 7.77 (12/15) 6.24 7.55 7.89
1hoe 74 β 9.39 (5/13) 9.87 (16/35) 7.96 10.00 10.37
1ezgA 82 β 11.03 (40/44) 10.42 (42/66) 9.66 10.35 10.62
1sfp 111 β 7.48 (2/18) 11.07 (5/11) 9.32 11.09 11.59
1b2pA 119 β 12.52 (31/56) 10.01 (18/25) 8.76 10.89 11.32

Columns 1–3 list the PDB code, size, and type of the test proteins. Column “Best Cluster” lists the RMSDs of the representative decoys of the best clusters. In this column, the first number in parentheses denotes the rank of the best cluster and the second number is the total number of clusters. Column “best” lists the RMSDs of the best decoys. Columns “1%” and “2%” list the average RMSDs of the top 1% and 2% decoys, respectively. The results of TOUCHSTONE-II are from Zhang et al. (2003).