A systematic approach for comprehensive T-cell epitope discovery using peptide libraries.

T-cell response to peptides bound on MHC Class I or Class II molecules is essential for immune recognition of pathogens. T-cells are activated by specific peptide epitopes that are determined within the antigen processing pathways and presented on the surface of other cells bound to MHC molecules. To determine which part of allergenic or pathogenic proteins can stimulate T-cells is important for the treatment of diseases. We sought to take advantage of the falling cost of synthetic, screening grade peptides, and devise a comprehensive, non-hypothesis-driven screen for T-cell epitopes. We were interested in the study of celiac disease (CD) and used the ELISPOT technique to perform a large number of T-cell assays. We therefore needed to compensate for the lack of statistical data analysis methods for ELISPOT assays. We describe a method to comprehensively screen for T-cell epitopes within a family or a group of proteins. We have implemented an algorithm to generate a set of unique short peptide sequences that incorporate all possible epitopes within a group of proteins. T-cell assays were performed in 96-well plates using the ELISPOT assay to screen for responses in CD patients against any epitopes in glutens. We describe a statistical model to fit the data and an Expectation Maximization algorithm to estimate the parameters of interest and analyze the resulting data. Implementations of our algorithms in R or Perl are available at http://bioinf.wehi.edu.au/folders/immunol.