In recent years, there has been an increase in the number of reported cases of infection in various animals, which can interfere with the ante-mortem diagnosis of animal tuberculosis caused by . In this study, whole genome sequencing (WGS) was used to search for protein-coding genes to distinguish from . In addition, the population structure of the available genomic WGS datasets is described, including three novel Belgian isolates from infections in alpacas. Candidate genes were identified by examining the presence of the regions of difference and by a pan-genome analysis of the available WGS data. A total of 80 genes showed presence-absence variation between the two species, including genes encoding Proline-Glutamate (PE), Proline-Proline-Glutamate (PPE), and Polymorphic GC-Rich Sequence (PE-PGRS) proteins involved in virulence and host interaction. Filtering based on predicted subcellular localization, sequence homology and predicted antigenicity resulted in 28 proteins out of 80 that were predicted to be potential antigens. As synthetic peptides are less costly and variable than recombinant proteins, an approach was performed to identify linear and discontinuous B-cell epitopes in the selected proteins. From the 28 proteins, 157 B-cell epitope-based peptides were identified that discriminated between and species. Although confirmation by testing is still required, these candidate synthetic peptides containing B-cell epitopes could potentially be used in serological tests to differentiate cases of from infection, thus reducing misdiagnosis in animal tuberculosis surveillance.