The lack of appropriate methods for subclinical diagnosis hampers the control and eradication of Johne's disease in sheep. As infection is insidious with prolonged incubation, it is usually difficult to assess performance of diagnostic methods during progressive stages of natural infection. In the present study, the performance of various tests were therefore evaluated on 15 experimentally infected sheep and 8 in-contact sheep. These animals had various grades of infection (negative, suspected, mild, moderate and severe) categorised on the basis of a set of diagnostic methods (faecal smear, culture and PCR, tissue PCR, ELISA immunoperoxidase and histopathology) over a period of 390 days. Four animals each from the infected and incontact groups were either in the negative or the suspected infection categories, and hence were not included in the assessment. Thus, the performance of the diagnostic methods was assessed on 15 sheep (11 from infected and 4 from incontact groups). Among 15 infected sheep, 9 (67.7%) were in the mild and 3 (33.3%) each were in the moderate and the severe infection categories. The sensitivities of the clinical diagnostic methods were 80% for I/D johnin test, over 73% for faecal smear and faecal PCR, 60% ELISA, over 53% for faecal culture and AGID, and 40% for blood PCR. Among post-mortem diagnostic methods, the highest sensitivity was obtained for immunoperoxidase test (80%), followed by tissue PCR (~67%) and acid-fast bacilli demonstration in tissues (40%). The study revealed that among antemortem tests, johnin test followed by feacal PCR, and among postmortem tests, immunoperoxidase were the sensitive test for the diagnosis of paratuberculosis. The performance of these tests in relation to severity of infection and its implication in the diagnosis and control of paratuberculosis has been discussed.