Vaccination against Mycobacterium avium subspecies paratuberculosis (Mptb) infection is effective as it prevents 90% of clinical cases. However, it does not prevent infection, and some vaccinates succumb to the disease while others can spread the infection through the shedding of large numbers of acid fast bacilli in their faeces. Importantly, current commercial vaccines cause occasionally severe and commonly persistent injection-site lesions in sheep and are hazardous to humans in cases of accidental self-injection. While there is ample justification for developing better vaccines for paratuberculosis, the incentive to do so is limited because lengthy, and therefore expensive, efficacy trials are required for this chronic disease. To overcome this we devised a cost effective strategy for candidate vaccine evaluation through short term trials using a defined infection model. Correlates of protection were identified using early immune parameters that predict eventual disease outcome. By evaluating specific parameters of an animal’s early phenotype, measured using immunological and transcriptomic approaches, we are able to assess whether novel vaccine candidates are likely to be effective in protecting against disease. In addition, we evaluated the effect of adjuvants and vaccine formulations on the development of injection-site lesions and identified those that cause no or merely transient lesions. We screened potential adjuvant-antigen combinations to drive the host immune response towards protection against this disease. Using knowledge gained from experimental trials in sheep, we identified a safe candidate vaccine that apparently stops faecal shedding and certainly decreases disease severity in cattle.