Understanding how viruses offer protection against closely related emerging strains is vital for creating effective vaccines. For many viruses, multiple serotypes often co-circulate and testing large numbers of vaccines can be infeasible. Therefore the development of an in silico predictor of cross-protection between strains is important to help optimise vaccine choice. Here we present a sparse hierarchical Bayesian model for detecting relevant antigenic sites in virus evolution (SABRE) which can account for the experimental variability in the data and predict antigenic variability. The method uses spike and slab priors to identify sites in the viral protein which are important for the neutralisation of the virus. Using the SABRE method we are able to identify a number of key antigenic sites within several viruses, as well as providing estimates of significant changes in the evolutionary history of the serotypes. We show how our method outperforms alternative established methods; standard mixed effects models, the mixed effects LASSO, and the mixed effects elastic nets. We also propose novel proposal mechanisms for the Markov chain Monte Carlo simulations, which improve mixing and convergence over that of the established component-wise Gibbs sampler.