Skip to main content

Table 2 Performance comparison of different methods using five-fold cross-validation for different datasets.

From: Prediction of B-cell epitopes using evolutionary information and propensity scales

Dataset Method AUC ACC SEN SPE MCC PPV
Sollner BEEPro 0.9987 0.9929 0.9604 0.9946 0.9281 0.9042
AntiJen#1 BEEPro 0.9930 0.9731 0.9680 0.9735 0.8491 0.7688
AntiJen#2 BEEPro 0.9907 0.9580 0.9700 0.9562 0.8402 0.7668
  LEPS NA 0.7381 0.2672 0.8448 0.1010 0.2885
  BepiPred NA 0.5552 0.5179 0.5761 0.0604 0.2202
  ABCPred 0.8 NA 0.4470 0.6733 0.4040 0.0546 0.2183
  BCPred NA 0.5392 0.5884 0.5487 0.0893 0.2334
  FBCPred NA 0.5145 0.6031 0.5121 0.0673 0.2233
HIV BEEPro 0.9907 0.9454 0.9490 0.9433 0.8853 0.9098
  LEPS NA 0.6345 0.4833 0.7484 0.2276 0.7144
  BepiPred 0.6000 0.5672 0.5016 0.6085 0.0972 0.6122
  ABCPred 0.7 NA 0.5659 0.8797 0.1465 0.0564 0.5633
  BCPred NA 0.6657 0.8018 0.5457 0.2980 0.6555
  FBCPred NA 0.6713 0.7320 0.5820 0.2781 0.6556
Pellequer BEEPro 0.9874 0.9373 0.9256 0.9435 0.8621 0.8935
  BepiPred 0.6710 NA NA NA NA NA
PC BEEPro 0.9950 0.9550 0.9708 0.9468 0.9036 0.9058
  LEPS NA 0.6166 0.1278 0.8833 0.0365 0.4512
  BepiPred NA 0.5533 0.4823 0.5972 0.0749 0.3819
  ABCPred 0.8 NA 0.4889 0.6546 0.4026 0.0513 0.3621
  BCPred NA 0.5283 0.5092 0.5935 0.0443 0.3607
  FBCPred NA 0.5220 0.5103 0.5255 0.0317 0.3526
Benchmark BEEPro 0.9100 0.9200 0.7100 0.9400 0.5700 0.5200
  CBTOPE 0.8900 0.8400 0.8000 0.8500 NA 0.3100
  DiscoTope 0.6000 0.7500 0.4200 0.7900 NA 0.1600
  CEP 0.5400 0.7400 0.3100 0.7800 NA 0.1100
  ClusPro(DOT) best model 0.6900 0.8900 0.4500 0.9300 NA 0.3900
  Patch Dock best model 0.6600 0.8500 0.4300 0.8900 NA 0.2600
  PSI-PRED best patch 0.6000 0.8200 0.3300 0.8600 NA 0.1900
  ProMate 0.5100 0.8400 0.0900 0.9200 NA 0.1000
  1. 1LEPS, BepiPred, ABCPred, BCPred, FBCPred performances were previously compiled by Wang et al. [9]
  2. 2CBTOPE, DiscoTope, CEP, ClusPro, Patch Dock, PSI-PRED, ProMate performances were previously compiled by Ansari and Raghava [13]