Promotionsvortrag Physik: A bottom-up approach towards understanding resistance evolution

Ankündigung des Promotionsvortrags von: Herrn Serhii Aif

principles increasingly inform therapy design, they rarely account for spatial dynamics: in dense cellular populations, spatial organization and physical cell-cell interactions strongly alter which mutations survive.

In the first part of this talk, I present experimental and computational results using a genetically engineered microbial model to track resistant clones carrying a fitness cost. During range expan-sion, radial growth and selection against costly mutations reach a quasi-stable equilibrium, termed inflation-selection balance. Here, collective cell pushing maintains disadvantaged clones at the colony front, increasing their likelihood of acquiring compensatory mutations. Agent-based simu-lations demonstrate that localized growth constraints and cell-cell pushing are sufficient to drive this mechanism.

In the second part, I introduce Reinforcement Failing, a computational framework using diverse reinforcement learning agents to identify physical constraints that cause simplified therapy models to break down. Applied to adaptive cancer therapy, accounting for these spatial dynamics pro-duces treatment strategies that generalize significantly better to complex tumor behavior.

Together, this work demonstrates how combining physical cell-cell interactions with reinforcement learning can uncover novel biological insights and support the rational design of next-generation therapies.

(Vortrag auf Englisch)

Dem Vortrag schließt sich eine Diskussion von 15 Minuten an. Vortrag und Diskussion sind öffentlich. Diesen Verfahrensteilen folgt ein nicht öffentliches Rigorosum von 45 Minuten.