Table 1.
Score from ROC curves and cumulative distributions distinguish structural alignment methods
Score with methoda | |||||||
---|---|---|---|---|---|---|---|
MEASURE | SSAP | STRUCTAL | DALI | LSQMAN | CE | SSM | Best-of-All |
ROC area (native score) | 0.83 | 0.94 | 0.94 | 0.70 | 0.89 | 0.80 | NAb |
ROC area (SAS score) | 0.91 | 0.93 | 0.94 | 0.87 | 0.94 | 0.93 | 0.93 |
ROC area (GSAS score) | 0.91 | 0.94 | 0.94 | 0.84 | 0.94 | – | 0.93 |
ROC area (SI score) | 0.80 | 0.79 | 0.84 | 0.72 | 0.81 | 0.79 | 0.76 |
ROC area (MI score) | 0.80 | 0.78 | 0.84 | 0.72 | 0.79 | 0.82 | 0.78 |
%TP at 1% FP (native score) | 9 | 42 | 50 | 0.4 | 29 | 3 | NA |
%TP at 1% FP (SAS score) | 48 | 32 | 47 | 29 | 46 | 25 | 22 |
%TP at 1% FP (GSAS score) | 48 | 30 | 47 | 20 | 46 | – | 15 |
%TP at 1% FP (SI score) | 16 | 2 | 16 | 4 | 9 | 0.8 | 1 |
%TP at 1% FP (MI score) | 16 | 3 | 16 | 7 | 10 | 4 | 3 |
% CAT–CAT at GSAS=5 Å | 19 | 68 | 37 | 29 | 50 | – | 74.0 |
% CAT–CAT at SAS=5 Å | 25 | 80 | 62 | 54 | 61 | 73 | 85.7 |
% CAT–CAT at SI=5 Å | 16 | 56 | 36 | 36 | 37 | 45 | 60.4 |
% CAT–CAT at MI=0.8 | 26 | 62 | 44 | 28 | 50 | 53 | 65.5 |
% All pairs at GSAS=5 Å | 0.6 | 5.6 | 1.5 | 2.6 | 2.4 | – | 8.5 |
% All pairs at SAS=5 Å | 0.8 | 9.4 | 3.5 | 5.6 | 3.5 | 7.1 | 13.5 |
% All pairs at SI=5 Å | 1.4 | 17.2 | 4.7 | 11.1 | 6.1 | 13.8 | 24.2 |
% All pairs at MI=0.8 | 2.6 | 21.7 | 7.0 | 6.9 | 12.2 | 13.6 | 25.2 |
A high value is always better. The best scores for each measure are shown in bold, as are the scores for Best-of-All when they are the best.
There is no native score for the Best-of-All method.