SEPPA: a computational server for spatial epitope prediction of protein antigens


In recent years, a lot of efforts have been made in conformational epitope prediction as antigen proteins usually bind antibodies with an assembly of sequentially discontinuous and structurally compact surface residues. Currently, only a few methods for spatial epitope prediction are available with focus on single residue propensity scales or continual segments clustering. In the method of SEPPA, a concept of 'unit patch of residue triangle' was introduced to better describe the local spatial context in protein surface. Besides that, SEPPA incorporated clustering coefficient to describe the spatial compactness of surface residues. Validated by independent testing datasets, SEPPA gave an average AUC value over 0.742 and produced a successful pick-up rate of 96.64%. Comparing with peers, SEPPA shows significant improvement over other popular methods like CEP, DiscoTope and BEpro. In addition, the threshold scores for certain accuracy, sensitivity and specificity are provided online to give the confidence level of the spatial epitope identification. The web server can be accessed at Batch query is supported.

DOI: 10.1093/nar/gkp417

Extracted Key Phrases

1 Figure or Table

Citations per Year

224 Citations

Semantic Scholar estimates that this publication has 224 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@inproceedings{Sun2009SEPPAAC, title={SEPPA: a computational server for spatial epitope prediction of protein antigens}, author={Jing Sun and Di Wu and Tianlei Xu and Xiaojing Wang and Xiaolian Xu and Lin Tao and Y. X. Li and Zhi-Wei Cao}, booktitle={Nucleic Acids Research}, year={2009} }