Clinical microbiology is steadily undergoing a genetic revolution; instead of being cultured, clinical samples will instead be sequenced using whole-genome sequencing (WGS) technologies and the effectiveness of a panel of antibiotics inferred from the presence or absence of mutations in a set of known resistance genes. This has already happened in the UK; last March Public Health England switched to using WGS for routine diagnosis of tuberculosis. A major problem is that this approach cannot make a prediction if it encounters novel or rare mutations in the set of resistance genes. We will show how structural biology, combined with molecular simulation, is able to predict whether mutations confer resistance (or not) to (a) trimethoprim in S. aureus [1] and (b) rifampicin in M. tuberculosis (unpublished). Looking forward, this approach, or ones like it, could be extended to minimise the number of “escape routes” a protein has to abrogate the action of a novel antibiotic, which would help de-risk antibiotic drug development.