Skip to main content

Table 4 Prediction accuracies for an independent dataset

From: Predicting MHC class I epitopes in large datasets

  Allele Binders Non-Binders Total DynaPredPOS NetMHC SVMHC YKW
   (n) (n) (n) AUC AUC AUC AUC
1 A*0201 33 143 176 0.92 0.94 0.82 0.93
2 A*0301 11 165 176 0.77 0.92 0.70 0.85
3 A*1101 17 159 176 0.84 0.89 0.72 0.83
4 A*2402 37 139 176 0.90 0.78 0.90 0.64
5 B*0702 9 167 176 0.86 0.98 0.72 0.68
6 B*0801 10 166 176 0.97 0.92 0.97 0.61
7 B*1501 14 162 176 0.71 0.92 0.71 0.80
  Average AUC     0.85 0.90 0.79 0.76
  1. A set of 176 novel peptides, generated and tested by Lin et al [15], were used to test the prediction accuracy of the four methods in this study. The average AUC for each method is included at the bottom of each column.