Table 2.
Negative dataset | Method/tool | % SE | % SP | % ACC | MCC |
---|---|---|---|---|---|
IRPEP | BLAST | 50.50 | 99.20 | 74.85 | 0.57 |
BepiPred | 37.60 | 65.01 | 51.35 | 0.03 | |
LBtope | 42.21 | 76.34 | 59.32 | 0.20 | |
IBCE-EL | 77.80 | 14.91 | 46.26 | − 0.09 | |
INB | BLAST | 50.50 | 88.47 | 69.48 | 0.42 |
BepiPred | 37.60 | 66.60 | 52.14 | 0.04 | |
LBtope | 42.41 | 77.73 | 60.02 | 0.20 | |
IBCE-EL | 77.80 | 82.11 | 79.96 | 0.60 |
Table reports the sensitivity (% SE), specificity (% SP), accuracy (% ACC) and Matthew’s correlation coefficient (MMC) of the BLAST-based method, BepiPred, LBtope and IBCE-EL discriminating B cell epitopes in the BECIP dataset from non-B cell epitopes in two different datasets (IRPEP and INB). B cell epitope predictions with LBtope and IBCE-EL were carried out at the relevant web sites and BepiPred predictions were carried out using the standalone version of BepiPred (details in “Methods”).