The nature of the variability in the clinical and epidemiological consequences of Mycobacterium tuberculosis infection remains poorly understood. Environmental and host factors that contribute to the outcome of infection and disease presentation are well recognised, but the role of bacterial factors has been more elusive. The rapid increase in the understanding of the molecular basis of M tuberculosis over the past decades has revived research into its pathogenesis. DNA fingerprinting techniques have been used to distinguish between strains of M tuberculosis, and efforts to characterise the strains present within populations have led to increased understanding of their global distribution. This research has shown that in certain areas a small number of strains are causing a disproportionate number of cases of the disease. The sequencing of the complete genome of M tuberculosis has accelerated the development of molecular techniques to differentiate strains according to their genetic polymorphisms. Investigation into the reasons why some strains are predominant by genetic strain-typing techniques may clarify which bacterial factors contribute to disease. This knowledge has the potential to influence control and prevention strategies for tuberculosis in the future. However, there are still limitations in these techniques and their results. This review discusses molecular epidemiology and genetic studies, and their contribution to the understanding of the links between genotypic and phenotypic characteristics of M tuberculosis strains.