Inferring immune-driven selection in SARS-CoV-2 evolution

Understanding the effects of individual mutations during viral evolution is crucial for elucidating how viruses adapt. Mutations can alter viral transmissibility or enable immune escape—for example, certain substitutions in the receptor-binding domain (RBD) allow the virus to evade antibody binding. Recent population-genetic frameworks enable the inference of selection coefficients from evolutionary trajectories. Building on this foundation, we incorporate antibody-related terms to investigate immune-driven evolution. Deep mutational scanning (DMS) experiments provide large-scale measurements of how individual mutations affect antibody binding. Using existing DMS datasets, we obtained escape scores for thousands of mutations against a panel of antibodies. We then integrated these scores into our inference framework to jointly estimate both fitness and immune-escape effects. This approach provides a quantitative understanding of how immune pressure has shaped the evolutionary dynamics of SARS-CoV-2.