Antimicrobial resistance (AMR) is a pressing global health crisis that demands innovative evolutionary and ecological strategies. My research journey into this field stems from a foundational background in genotype-phenotype maps, which naturally led me to explore how we can understand, predict, and ultimately steer evolutionary trajectories. In the first part of this talk, I will briefly discuss my transition into the field of AMR and share recent insights into the mathematical modeling of antibiotic resistance.

The core of the presentation will focus on a promising, underexplored strategy to tackle AMR: sequential therapies that exploit collateral sensitivity (CS), a phenomenon where resistance to one drug induces susceptibility to another. Using a four-genotype stochastic birth-death model with two bacteriostatic antibiotics, we identify switching periods that maximize bacterial extinction under subinhibitory concentrations. Counterintuitively, we show that fast sequential therapies are suboptimal; allowing some resistance to evolve is actually a key ingredient for these therapies to succeed.

I will outline the geometric distribution framework we developed to accurately predict cumulative extinction probabilities, highlighting how extinction depends nonlinearly on switching periods with stepwise increases aligned to discrete switch events. Furthermore, we will explore the evolutionary trade-offs involved—such as the Pareto front between maximizing extinction and minimizing resistance—and the requirement for strong reciprocal CS. Ultimately, these results provide quantitative design principles for in vitro and clinical sequential antibiotic therapies, underscoring the immense potential of CS‐guided regimens to suppress resistance evolution and eradicate infections.