Abstract Staphylococcus aureus is a leading cause of infection worldwide and a major driver of antibiotic resistance. Although many experimental staphylococcal vaccines have been reported, all vaccines tested to date in human trials have failed for unclear reasons. Unlike mice, humans are exposed to S. aureus beginning early in life, leading to generation of antibodies to S. aureus antigens. In preliminary experiments, we have shown that select human anti-S. aureus antibodies are protective, but many are not. In mice exposed to S. aureus, vaccination against protective antigens leads to immunity against S. aureus whereas vaccination against non-protective antigens induced recall of non-protective immunity which further interferes with protective antibodies by direct competition. Based on these findings, we generated a model of how pre-existing antibodies shape vaccine responses and how this predicts novel ways to develop effective vaccines against S. aureus. To query the validity of our model, in Aim 1, we will recruit children and old adults with invasive S. aureus infections, survey the anti-S. aureus antibody profile and define functionally protective antibody responses to S. aureus antigens. In Aim 2, we will identify and characterize protective and non-protective anti-S. aureus antibodies from candidate samples acquired in Aim 1, and assess structural and functional features of the specific antibodies that confer protection or non-protection. In Aim 3, we will study these antibodies and their target in the context of naïve mice and mice previously exposed to S. aureus. We will evaluate mechanisms whereby non-protective memory shapes vaccine efficacy and test strategies that circumvent interference. Overall, using the novel model systems, we aim to develop a more predictive framework for explaining staphylococcal vaccine failures and developing novel strategies for effective vaccination against S. aureus.