Most TB disease, both drug-sensitive and drug-resistant, is recently transmitted via social interactions away from the household. Thus, there is a need to understand the adult and childhood social networks in South African townships that contribute to TB transmission.

Surrogate biomarkers are essential for public health interventions.

Transmission and infection need to be studied intensively.

A TB control program must be able to rapidly identify cases and commence therapy.

Operational research and implementation science are needed to enhance feasible and available infection control strategies and support the development and testing of new approaches.

Strains of TB are more diverse than previously appreciated.

The immense amount of variation occurring between strains suggests that purifying selection is reduced in Mycobacterium tuberculosis (M.tb.).

Strain variation is extensive within individual study sites and across regions, with the University of Stellenbosch group reporting more than 150 strains per square kilometer in the study areas.

The population structure of M.tb. is highly variable, with a dominance of Beijing genotype (both typical and atypical) being uncovered in some study areas and playing a major role in driving drug resistance in South Africa. Intriguingly, Beijing strain has not (yet) been observed in Zambia, and the KZN strain predominates in XDR TB cases in that area.

DNA sequence data are highly informative and will soon be the standard for molecular epidemiology.

Although genetic and molecular epidemiology is powerful, it has been studied only in some areas of the region.

While new diagnostics could detect 400,000 cases each year, the realities of using new versus current diagnostics need to be taken into account.