In silico veritas? Potential limitations for SARSCoV-2 vaccine development based on T-cell epitope prediction

Abstract

The development of a vaccine able to prevent infection or severe disease course of SARS-CoV-2 is a priority to stem the current COVID-19 pandemic and to be better prepared for future flare-ups. To accelerate T-cell immunogen design, many current approaches are employing epitope prediction strategies. Although such approaches have great merit, it is also important that unbiased approaches to characterizing the T-cell response to SARS-CoV-2 are incorporated into vaccine design, in order to generate a comprehensive picture of the total virus-specific T-cell response and to define correlates of protective immunity against the virus. Ever since the first identification of binding motifs for T-cell antigens presented by HLA class I molecules by Rammensee and colleagues almost 30 years ago, epitope identification has been greatly facilitated by epitope prediction algorithms [1]. Over the years, many vaccines designs targeting infectious pathogens as well as cancer neoantigens have been based on in silico prediction of potential HLA class I-restricted epitopes, and a series of vaccine candidates that apply such strategies to SARS-CoV-2 are currently in development. However, even though numerous prediction algorithms have been developed and gradually improved, there are several considerations that may threaten or limit the success of such approaches.