BCEPS: A Web Server to Predict Linear B Cell Epitopes with Enhanced Immunogenicity and Cross-Reactivity

Abstract

Prediction of linear B cell epitopes is of interest for the production of antigen-specific antibodies and the design of peptide-based vaccines. Here, we present BCEPS, a web server for predicting linear B cell epitopes tailored to select epitopes that are immunogenic and capable of inducing cross-reactive antibodies with native antigens. BCEPS implements various machine learning models trained on a dataset including 555 linearized conformational B cell epitopes that were mined from antibody–antigen protein structures. The best performing model, based on a support vector machine, reached an accuracy of 75.38% ± 5.02. In an independent dataset consisting of B cell epitopes retrieved from the Immune Epitope Database (IEDB), this model achieved an accuracy of 67.05%. In BCEPS, predicted epitopes can be ranked according to properties such as flexibility, accessibility and hydrophilicity, and with regard to immunogenicity, as judged by their predicted presentation by MHC II molecules. BCEPS also detects if predicted epitopes are located in ectodomains of membrane proteins and if they possess N-glycosylation sites hindering antibody recognition. Finally, we exemplified the use of BCEPS in the SARS-CoV-2 Spike protein, showing that it can identify B cell epitopes targeted by neutralizing antibodies.

Prediction of B cell epitopes in proteins using a novel sequence similarity-based method

“…Unlike competing tools, BepiBlast can also report if predicted B cell epitopes come from BLAST hits to neutralizing B cell epitopes as well as the accessibility and flexibility of B cell epitopes. Antibodies generated against predicted linear B cell epitopes do often fail to recognize the native protein, but this can be compensated by selecting B cell epitopes with enhanced flexibility and solvent accessibility 18 , 38 , 43 , 44 .…”

“…Given the essentially endless diversity of the BCR and antibodies, almost any peptide can be suitable for recognition and hence be a B cell epitope. Therefore, the most complex and recent B-cell prediction www.nature.com/scientificreports/ methods make use of machine learning (ML)-based models that are generated by training ML algorithms on feature data drawn from experimentally determined B cell epitopes and assumed non-B cell epitopes [14][15][16][17][18][19][20] . As a result of training, ML-algorithms capture subtle patterns into a single model that serve to distinguish B cell epitopes from non-B cell epitopes.…”

“…The first of such scales was introduced by Hopp and Woods 8 and many other scales followed latter, including those based on flexibility 9 , hydrophobicity 10 , 11 , surface accessibility 12 and antigenicity 13 . Most recent approaches to predict B cell epitopes use machine learning algorithms such as neural network 14 , 15 , support vector machine 16 – 18 or random forest 19 , 20 , which are trained on features of known B cell epitopes. As shown by various benchmark evaluations 21 – 23 , the performance of B cell epitope prediction methods can be quite low and there is still room for improvement.…”

In silico design of recombinant multi-epitope vaccine against influenza A virus

“…Ectodomain location, glycosylation sites, and solvent-accessible regions were predicted for the multi-epitope vaccine using the BCEPS web server ( http://imbio.med.ucm.es/bceps/)and NetSurfP ( https://services.healthtech.dtu.dk/service.php?NetSurfP-1.1 ) used to evaluate solvent accessible regions for selected B-cell epitopes. The NetSurfP server measures the solvent accessible regions of all amino acids in each selected B-cell epitope [ 50 ]. The parameters for BCEPS web server were set to default, which means the model was set to SVM; the number of aa was set at 16; the threshold was set to 0.5 and the immunogenicity was set for considering both CD4 and any human.…”

Quantitative In Silico Evaluation of Allergenic Proteins from Anacardium occidentale, Carya illinoinensis, Juglans regia and Pistacia vera and Their Epitopes as Precursors of Bioactive Peptides

“…In addition, the amino acid sequence and conformational similarities of these proteins contribute to their potential to cross-react with IgE cognate to allergens from other tree nuts or peanuts [ 34 , 35 , 36 ]. Contemporary strategies of epitope prediction assume that they can be discriminated from the entire protein chain due to amino acid composition, amino acid order, or oligopeptide composition [ 37 , 38 , 39 , 40 ]. Alternatively, it may be that that epitopes differ from entire protein chains in frequency of occurrence of bioactive fragments.…”