Tuberculosis (TB) remains a major cause of mortality and morbidity worldwide. The current vaccine, BCG, is only partially effective against TB, and drug-resistant strains of Mycobacterium tuberculosis are emerging at an alarming rate. The development of new TB control strategies requires an understanding of how the pathogen adapts to life within the host, in particular the ability of M. tuberculosis to persist in a latent state and thus avoid immune clearance. We have shown that specific M. tuberculosis virulence factors can impact on the development of the host immune response, and deletion of such virulence-encoding genes results in highly protective TB vaccine strains. We have also used genetic screens to identify M. tuberculosis components highly expressed during infection, in order to identify potential vaccine and drug targets. When tested in animal models, novel vaccines incorporating host-expressed proteins can serve to control the growth of M. tuberculosis at different stages of the pathogen life cycle. This includes enhanced clearance of bacteria at both the active and chronic states of M. tuberculosis infection, reduced bacterial load when delivered as a post-exposure treatment, and a marked ability of these ‘multistage’ vaccines to block reactivation of disease. Our current focus is to prepare the most promising vaccines for assessment of efficacy against TB in human trials.