LBEEP

LBEEP predicts linear B-cell epitopes from protein sequences using a Dipeptide Deviation from Expected Mean (DDE) amino acid composition descriptor to support vaccine design and related immunoinformatics applications.

Key Features:

• Dipeptide Deviation from Expected Mean (DDE): An amino acid composition-based feature descriptor (DDE) used to distinguish linear B-cell epitopes from non-epitopes.
• Exact epitope focus: Methodology emphasizes exact linear B-cell epitopes and non-epitopes rather than broader epitope-containing regions.
• Machine learning classifiers: Uses Support Vector Machine (SVM) and AdaBoost‑Random Forest for classification.
• Validation strategy: Performance assessed with five-fold cross-validation on error-free datasets and datasets from other studies, reporting overall accuracy between 61% and 73% with balanced sensitivity and specificity.
• Comparative performance: DDE feature vector improved accuracy compared to other amino acid-derived features by approximately 2% to 12%.
• Implementation: Implemented as a Perl-based application.

Scientific Applications:

• Vaccine design: Identification of linear B-cell epitopes to inform vaccine antigen selection.
• Synthetic vaccine design: Support for designing synthetic peptide-based vaccine candidates via epitope prediction.
• Systems therapeutics: Contribution to target identification in therapeutic development workflows.
• Diagnostics and molecular target identification: Prioritization of linear B-cell epitopes for diagnostic marker and therapeutic target discovery.

Methodology:

Compute DDE feature vectors from peptide sequences, train and evaluate classifiers using Support Vector Machine and AdaBoost‑Random Forest with five-fold cross-validation on error-free and external datasets, and compare accuracy against other amino acid-derived feature representations (reported accuracy 61%–73%; improvement 2%–12%).