Staged heterogeneity learning to identify conformational B-cell epitopes from antigen sequences

BMC Genomics

Table 2 Prediction results by straightforward staged learning

Predictor	Dataset	Recall	Precision	F-score
SePre.v ₀	Exp104	0.793	0.085	0.153
SePre.v ₀	Exp163	0.744	0.055	0.102
SePre.v ₀	Ren92unbound	0.923	0.186	0.310

SePre.v ₀ in this table is different from the proposed SePre model. Its second-stage model was trained on the leave-one-out cross-validation (LOOCV) results from the first-stage model. In LOOCV, if an antigen has multiple chains, these chains are all left out as tests in a round to reduce the coupling between a target antigen and the training data

ISSN: 1471-2164