Comprehensive Benchmarking of T-cell epitope prediction tools

Abstract

Abstract Computational prediction tools are commonly used to select T cell epitope candidates in several immunological applications including vaccine design, cancer neoantigen identification, development of diagnostics and removal of unwanted immune responses against protein therapeutics. Most of these prediction tools are based on machine learning algorithms trained on MHC binding data or naturally processed MHC ligand data. The ability of currently available tools to predict T cell epitopes has not been comprehensively evaluated. In this study, we use a recently published dataset that systematically defined T cell epitopes recognized in vaccinia virus infected mice to evaluate the performance of all currently publicly available computational T cell epitope prediction tools. Using the same data set across all prediction methods provides an unbiased performance evaluation. The derived performance metrics allow immunologists to rationally decide on which prediction method to apply for their work at hand.