An entire genome sequence will provide valuable information, but the genome of only one individual will limit interpretation of that information. Knowledge concerning genome variation in both eukaryotic and prokaryotic organisms such as Mycobacterium tuberculosis is likely to yield information of equal value and provide fundamental insights concerning the function of the genome. The variability in the genome between individual strains may be small and well defined, but it may cause large phenotypic changes (e.g. point mutations causing drug resistance). Clinical and epidemiological observations have led to the development of hypotheses, assumptions and models concerning disease dynamics. However, genome variation studied by molecular epidemiology has made new insights possible, which have allowed us to examine prevailing dogmas concerning tuberculosis. Recent results suggest that historical dogmas may well hold true in some communities, but not all. The information gathered from studying strain variation can be used for modelling disease dynamics, prediction of epidemics, policy planning and for monitoring the outcome of new interventions, as well as for gaining insight into the life processes of the organism. However, molecular epidemiology has its own limitations, some of which result from our lack of understanding of genome variation. We need further information in order to understand clonality and evolution of this organism so that our use of molecular tools in epidemiology and drug development may become more relevant and accurate.