AutoEpiCollect

AutoEpiCollect generates safe and immunogenic epitopes from missense mutations in oncogenes to streamline preclinical vaccine design.

Key Features:

• Machine learning-driven epitope ranking: The software incorporates a probabilistic logistic regression model trained on experimental T-cell assay data to rank epitopes.
• MHC Class I and Class II epitope prediction: Generates both MHC Class I epitope/HLA pairs and MHC Class II epitope/HLA pairs from input point mutations.
• Population coverage optimization: Employs PCOptim and PCOptim-CD to streamline epitope lists and attain optimized vaccine population coverage.

Scientific Applications:

• Pan-cancer vaccine design: Applied to design a pan-cancer vaccine targeting missense mutations in PIK3CA, the proto-oncogene encoding the p110ɑ catalytic subunit of the PI3K kinase protein.
• Epitope-HLA-TCR complex modeling: Predicted immunogenic epitopes can be used to construct 3D models of peptide-HLA and peptide-HLA-TCR complexes to analyze epitope binding potential and TCR interactions.

Methodology:

AutoEpiCollect uses a probabilistic logistic regression model trained on experimental T-cell assay data. Users input point mutations from an oncogene of interest, and the software outputs ranked MHC Class I and Class II epitope/HLA pairs. Population coverage is optimized using PCOptim and PCOptim-CD.