BCIgEPred

BCIgEPred predicts linear B-cell IgE epitopes in allergenic proteins to identify exact antigenic determinants involved in IgE-mediated allergic responses.

Key Features:

• Focus on Linear Epitopes: Targets exact linear B-cell IgE epitopes rather than broader epitope-containing regions, which is pertinent for allergens affected by processing or digestion.
• Dataset Utilization: Uses a dataset comprising experimentally verified exact IgE, IgG, IgM, and IgA epitopes for model training and evaluation.
• Machine Learning Models: Implements Support Vector Machine (SVM) and Random Forest (RF) classifiers using a Dipeptide Deviation from the Expected mean (DDE) feature vector derived from sequence data.
• Validation and Performance: Validated by five-fold cross-validation and independent dataset testing, reporting balanced accuracy of 74–78%, area under the ROC curve >0.8, and accuracy improvements of 16–28% over existing methods.
• Implementation and Integration: Developed as a Perl-based framework that can operate standalone or be integrated into broader allergen prediction workflows.

Scientific Applications:

• Allergy diagnostics and research: Identifies potential allergenic proteins and exact IgE epitopes to inform diagnostics and studies of IgE-mediated responses, including food allergies.

Methodology:

Perl-based implementation that computes DDE (Dipeptide Deviation from the Expected mean) feature vectors from sequence data and trains SVM and RF models, with performance assessed by five-fold cross-validation and independent dataset testing using experimentally verified exact IgE, IgG, IgM, and IgA epitopes.