TROLLOPE

TROLLOPE predicts linear T-cell epitopes of Hepatitis C virus (TCE-HCV) from protein sequence data to accelerate epitope discovery for vaccine and therapeutic research.

Key Features:

• Diverse Feature Descriptors: Employs 12 sequence-based feature descriptors encompassing physicochemical properties, composition-transition-distribution information, and simple composition data.
• Machine Learning Integration: Uses 12 machine learning algorithms combined with the feature descriptors to construct 144 base-classifiers.
• Feature Selection Strategy: Implements a feature selection process to identify the most promising base-classifiers for inclusion in the final model.
• Meta-Classifier Development: Integrates selected base-classifiers into a stacked meta-classifier to synthesize classifier strengths and improve predictive performance.
• Validation and Performance: Validated by cross-validation and independent testing, achieving reported accuracies of 0.745 (cross-validation) and 0.747 (independent testing).

Scientific Applications:

• HCV epitope discovery: Enables large-scale prediction of linear T-cell epitopes (TCE-HCV) from sequence data for downstream experimental validation.
• Vaccine and therapeutic research: Supports prioritization of candidate epitopes for HCV vaccine design and immunogen selection to improve immune-targeting strategies.

Methodology:

Integrates 12 sequence-based feature descriptors, constructs 144 base-classifiers using 12 machine learning algorithms, applies feature selection to choose effective base-classifiers, combines them into a stacked meta-classifier, and evaluates performance via cross-validation and independent testing.