B cell epitopes prediction

The B cell epitopes were predicted by the seven tools embedded in the Immune Epitope Database and Analysis Resource (IEDB). More specifically, BepiPred-1.0 , BepiPred-2.0 , Chou and Fasman beta turn prediction, Emini surface accessibility scale, Karplus and Schulz flexibility scale, Kolaskar and Tongaonkar antigenicity scale and Parker hydrophilicity prediction are used for predicting sites belonging to B cell epitopes. The parameters are all set on default. The sites confirmed by at least five tools would be labeled in red. The thresholds of each tool are listed as the table right. In this database, only amino acids that be confirmed by at least five tools would be part of B cell epitopes.

Bepipred >0.35
Chou & Fasman>average
Emini  >1
Karplus & Schulz   >average
Kolaskar & Tongaonkar   >average
Parker >average

T cell epitopes prediction

The T cell epitopes prediction were divided into two parts. One of them are presented by human leukocyte antigen (HLA) I allele and would induce the activation of CD8 T cells. This type of T cell epitopes were predicted by NetMHCpan 4.0, MHCflurry and DeepHLApan. Another type of T cell epitopes presented by HLA II alleles and induce the activation of CD4 T cells were predicted by MixMHC2pred and NetMHCIIpan. The thresholds to defined potential T cell epitopes of each tool are listed in the table right. In the prediction of T cell epitopes presented by HLA I alleles, all peptides with length range from 8 to 11 were selected and combined with previous HLA I alleles to create HLA-peptide pairs. It’s similar to predict that presented by HLA II alleles, with the difference that peptide length range from 15 to 28. Only HLA-peptide pairs satisfied with all thresholds of used tools would be potential T cell epitopes in this study. The HLA alleles with their frequency in Han Chinese ancestry used for T cell epitopes prediction can be download in the following link: HLA alleles file

DeepHLApan binding score>0.5
DeepHLApan immunogenic score>0.5
NetMHCpan affinity<500nM
NetMHCpan Rank<2%
MHCflurry Rank<2%
NetMHCIIpan Rank<10%


E-mail:        zhanzhou@zju.edu.cn

Address:      College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, Zhejiang

Citation: Wu J, Chen W, Zhou J, Zhao W,
Sun Y, Zhu H, Yao P, Chen S, Jiang J * and Zhou Z * (2020) COVIEdb: A Database for Potential Immune Epitopes of Coronaviruses.
Front. Pharmacol. 11:572249. doi: 10.3389/fphar.2020.572249