Skip to main content

Table 6 The area under the ROC curves for classification of near-native and non-native model

From: ROTAS: a rotamer-dependent, atomic statistical potential for assessment and prediction of protein structures

  Targets <|P|> <|N|> dDFIRE OPUS_PSP RWplus GOAP ROTAS
4state_reduced 7 195 468 0.86 0.80 0.81 0.91 0.92
fisa 2 47 453 0.79 0.60 0.79 0.79 0.77
lmds 2 60 439 0.74 0.64 0.66 0.61 0.56
fisa_casp3 2 20 1672 0.74 0.58 0.72 0.68 0.70
moulder 19 151 169 0.95 0.93 0.95 0.95 0.96
rosetta 27 50 50 0.71 0.66 0.74 0.75 0.77
I-TASSER 31 229 217 0.79 0.71 0.77 0.80 0.80
Amber99 41 219 821 0.87 0.79 0.83 0.93 0.93
CASP5-8 89 14 7 0.82 0.75 0.84 0.83 0.84
Total 220 105 245 0.82 0.75 0.82 0.84 0.85
p-value     1.02E-04 7.70E-27 1.30E-03 2.03E-06  
  1. <|P|>: Averaged number of positive (near-native) models in each target.
  2. <|N|>: Averaged number of negative (non-native) models in each target.
  3. p-value: P value of paired t-test of the difference of the AUC between ROTAS and the given potential.
  4. The best scores are highlighted in bold type.