CD8TCEI-EukPath

CD8TCEI-EukPath predicts antigen epitopes derived from eukaryotic pathogens that can stimulate host CD8+ T cells to support epitope-based vaccine discovery and immunological research.

Key Features:

• Predictive Capability: Screens peptide-coding sequences to identify epitopes from eukaryotic pathogens that can trigger adaptive CD8+ T-cell responses.
• Feature Representation: Integrates multiple amino acid sequence-based hybrid features as input for epitope classification.
• Feature Selection: Applies a feature selection technique to identify 520 optimal hybrid features for modeling.
• Classification Algorithm: Models selected features using the LightGBM machine learning algorithm for epitope prediction.
• Performance Metrics: Achieves 79.255% accuracy in ten-fold cross-validation and 78.169% accuracy on independent tests distinguishing CD8+ T-cell epitopes (TCEs) from non-CD8+ TCEs.

Scientific Applications:

• Vaccine Discovery: Enables rapid screening of pathogen-derived peptides to prioritize epitope-based vaccine candidates against infectious eukaryotes such as parasitic protozoans.
• Immunological Research: Provides epitope predictions to support studies of host-pathogen interactions and CD8+ T-cell-mediated adaptive immunity.

Methodology:

Integrates multiple amino acid sequence-based hybrid features, applies feature selection to obtain 520 optimal features, trains a LightGBM classifier, and evaluates performance with ten-fold cross-validation and independent testing.