Host gene expression profiling is a widely used research tool for assessing the host response to infection in order to provide insight into the immunopathophysiology of disease, as well as the analysis of disease progression and treatment response. It has recently been applied for the diagnosis of tuberculosis in children in Africa, as a result of the implementation of novel statistical methodology that enabled the reduction of a large number of significantly differentially expressed genes into a minimal set, and the development of a ‘disease risk score’ that could be used to develop a diagnostic test. Whilst the experimental and statistical methodologies are now in place to generate minimal transcriptional signatures that can distinguish disease states, the challenge is how to take these forward into development of a diagnostic test for use in clinical resource-poor settings.