HavePhAIth: Human phage therapy against the ESKAPE pathogens using AI

Human Phage Therapy (PT) is a promising route for the treatment of drug-resistant bacterial infections. Belgium is currently leading the implementation of PT in Europe, and the technique is currently in operations at the Queen Astrid Military Hospital (QAMH) and at UZ Leuven. A Multidisciplinary Phage Task Force (MPTF) has been set up within UZ Leuven to provide a PT framework for patients with difficult-to-treat infections. However, the current design strategies of phage cocktails are a black box, relying on empirical rules that fail to leverage the rapidly expanding omics data to predict bacteria-phage interactions. This in turn limits the inclusion criteria for patients to receive PT treatment. In my PhD research, I developed machine learning models of phage infectivity in P. aeruginosa that predict which phages from a collection can infect given bacterial strains based on their genomic content. As a member of the MPTF and collaborator of QAMH, two entities that will generate big datasets of omics/clinical data on PT in vivo, I will translate these modeling approaches to the ESKAPE pathogens in patients. This will put me in a unique position to assess the dynamics of bacteria-phage co-evolution in vivo, by inspecting longitudinal isolates from given patients undergoing treatment. Importantly, these analyses will enable us to extract design rules for phage  products, while productively translating our research towards "sur-mesure" phage treatment of individual patients.