Paratuberculosis, a chronic wasting disease in ruminants, is causing significant losses to both EU dairy and meat producers, due to a decreasing milk yield, loss of body weight and early replacements. The absence of adequate diagnostic tools for early detection of subclinically infected livestock severely interferes with animal welfare and effective eradication initiatives. Currently three strains of M. bovis and one MAP strain (K10) have been fully sequenced and deposited in NCBI genbank. We have sequenced a Danish clinical isolate, Ejlskov2007, which is not yet published. State of the art bioinformatics tools has been used to identify peptides in the MAP genome, which are predicted to bind to bovine MHC (BoLA) class II antigen presenting molecules. Comparative genomics tools has been used to further select MAP specific peptides 100% conserved in the two MAP strains and with low similarity to peptides from any of the M. bovis strains according to available sequence data. Unique MAP specific predicted epitopes is selected and is to be synthesized as 15–20 aa peptides. In addition we will select peptides specifically from antigens, which have previously shown to induce a CD4+ T cell response, again with the emphasis of avoiding similarities to M. bovis sequences. Blood from cattle experimentally infected with MAP will be used for proliferation and cytokine assays to determine if and which of the selected peptides that will be recognized specifically by CD4+ T cells from infected cattle. The goal is to combine positive responding peptides with new promising adjuvants in order to develop effective bovine MAP vaccines which do not cross react with traditional skin tests for M. bovis infections.