In The Lancet HIV, Manon Ragonnet-Cronin and colleague present an analysis of HIV sequence data from the UK to describe “potential non-disclosed men who have sex with men” (pnMSM). Their analysis is an illustration of the tremendous power of molecular analysis tools, applied to a powerful clinical dataset, generated through routine HIV resistance testing in clinical care settings. In the study, data about the relationship of viruses to one another were paired with data on sexual identity to identify men who were thought to be heterosexual, but whose HIV viruses clustered only with the viruses of men. It is important to discuss the limitations of data such as those analysed here, to be explicit about where such data can be useful in public health response, and to delineate which uses for such data exceed how the data can be meaningfully interpreted.