Rapid diagnosis and treatment are important for preventing transmission of Mycobacterium tuberculosis. However, the diagnosis of tuberculosis continues to pose serious problems, mainly because of difficulties in differentiating between patients with active tuberculosis and those with healed lesions, normal mycobacterium boris BCG (Bacillus Calmette Guerin) vaccinated individuals, and unvaccinated Manteux positives. Physicians still rely on conventional methods such as Ziehl-Neelsen (ZN) staining, fluorochrome staining, sputum culture, gastric lavage, and other non-traditional methods. Although the tuberculin test has aided in the diagnosis of tuberculosis for more than 85 years, its interpretation is difficult because sensitization with nontuberculous mycobacteria leads to false-positive tests. There have been numerous unsuccessful attempts to develop clinically useful serodiagnostic kits for tuberculosis. A number of proteinaceous and nonprotein antigens (such as acyltrehaloses and phenolglycolipids) have been explored from time to time for the development of such assays but they have not proved to be clinically useful. It has been difficult to develop an ELISA utilizing a suitable antigen because M. tuberculosis shares a large number of antigenic proteins with other microorganisms that may or may not be pathogenic. With the advent of molecular biology techniques, there have been significant advances in nucleic acid-based amplification and hybridization, which are helping to rectify existing flaws in the diagnosis of tuberculosis. The detection of mycobacterial DNA in clinical samples by polymerase chain reaction (PCR) is a promising approach for the rapid diagnosis of tuberculous infection. However, the PCR results must be corrected for the presence of inhibitors as well as for DNA contamination. In the modern era of genetics, marked by proteomics and genomics, the day is not far off when DNA chip-based hybridization assays will instantly reveal mycobacterial infections.