AimTo validate an optimized peptide-mediated magnetic separation (PMS)-phage assay for detection of viable Mycobacterium avium subsp. paratuberculosis (MAP) in milk. Methods and ResultsInclusivity, specificity and limit of detection 50% (LOD50) of the optimized PMS-phage assay were assessed. Plaques were obtained for all 43 MAP strains tested. Of 12 other Mycobacterium sp. tested, only Mycobacterium bovis BCG produced small numbers of plaques. LOD50 of the PMS-phage assay was 093 MAP cells per 50ml milk, which was better than both PMS-qPCR and PMS-culture. When individual milks (n=146) and bulk tank milk (BTM, n=22) obtained from Johne's affected herds were tested by the PMS-phage assay, viable MAP were detected in 31 (212%) of 146 individual milks and 13 (591%) of 22 BTM, with MAP numbers detected ranging from 6-948 plaque-forming-units per 50ml milk. PMS-qPCR and PMS-MGIT culture proved to be less sensitive tests than the PMS-phage assay. ConclusionsThe optimized PMS-phage assay is the most sensitive and specific method available for the detection of viable MAP in milk. Further work is needed to streamline the PMS-phage assay, because the assay's multistep format currently makes it unsuitable for adoption by the dairy industry as a screening test. Significance and Impact of the StudyThe inclusivity (ability to detect all MAP strains), specificity (ability to detect only MAP) and detection sensitivity (ability to detect low numbers of MAP) of the optimized PMS-phage assay have been comprehensively demonstrated for the first time.