Reinforcement Learning informs optimal treatment strategies to limit antibiotic resistance.

Drug resistant pathogens are a wide-spread and deadly phenomenon. Antimicrobial resistance was estimated to be associated with 4.95 million deaths worldwide in 2019. If resistance continues to develop at the current rate, bacterial infections are expected to surpass cancer as the leading cause of death worldwide by 2050. Despite this troubling trend, antimicrobial drug development has all but ceased. For the few new drugs that are approved, microbes develop rapid resistance through evolution by mutation and selection. Novel approaches to designing therapy that explicitly take into account the adaptive nature of microbial cell populations are desperately needed. Approaches that can design therapies given limited information about the evolving system are particularly important due to the limitations of clinical measurement. In this study, we explore a reinforcement learning (RL) approach capable of learning effective drug cycling policies in a system defined by empirically measured fitness landscapes. Given access to a panel of 15 β -lactam antibiotics with which to treat the simulated E. Coli population, we demonstrate that RL agents outperform two potential treatment paradigms at minimizing the population fitness over time. We also show that RL agents approach the performance of the optimal drug cycling policy. Crucially, we show that it is possible for RL agents to learn effective drug cycling protocols using current population fitness as the only training input. Our work represents a proof-of-concept for using AI to control complex evolutionary processes.