Antimicrobial resistance (AMR) is the ability of microorganisms, such as bacteria, to resist medicines like antibiotics. AMR is an escalating global health crisis that is already one of the biggest global killers. Increasingly, once easily treatable infections are becoming untreatable. Countering this threat requires new antibiotics that work against multidrug-resistant bacterial infections. However, developing antibiotics is complex, time-consuming and costly. Researchers must identify vulnerabilities within the most dangerous bacteria, then find molecules that interfere with critical functions, ultimately killing the bacteria.

GARDP and Google DeepMind have been working together on advanced modelling and analysis of several proteins identified as promising but unrealised targets for new antibiotics. These proteins are promising because they are important to certain bacteria, which the World Health Organization says critically require new treatments. If we identify a molecule that interferes with these proteins’ functions, it might become a powerful new antibiotic for the most dangerous multidrug-resistant bacteria.

The collaboration has enhanced GARDP’s understanding of the proteins and accelerated the timeline for assessing them as possible drug targets, bringing  us closer to developing new antibiotics. Furthermore, the work should enable GARDP to make greater use of AI technologies such as AlphaFold2. Google DeepMind provided analyses of the AlphaFold2-derived models, which can support future structure-based drug discovery efforts. It also produced a manual, enabling similar analyses of other proteins of interest.  Google DeepMind and GARDP are now exploring potential future collaborations.