Title: Large deviations theory of growing stochastic chemical reaction networks
Speaker: Dr. Praful Gagrani (University of Tokyo)
Host: Dr. Hong-Yan Shih (AS) / Dr. Wei-Hsiang Lin (AS)
Time: 11:00, August 20 (THU.), 2026
Place: Rm. 307, 3F, Cosmology Hall, NTU

Abstract:
Understanding the dynamics of growing chemical systems is essential for explaining the behavior of biological organisms. The metabolism of small organisms, such as bacteria, is often governed by autocatalytic reaction networks—systems in which the presence of certain species catalyzes the production of more of the same. In such organisms, the low copy numbers of key autocatalytic species make stochastic effects especially pronounced. Analyzing these stochastic fluctuations is therefore crucial for linking intracellular chemical dynamics to population-level behavior. A major challenge is that growing systems lack stationary distributions, which means that standard large deviations theory for stochastic processes does not directly apply. In this work, we extend large deviations principles to accommodate growing systems without stationary measures and derive analogues of classical results tailored to this setting. Our framework also yields estimators for stochastic growing systems that are better suited to growth-driven dynamics and easier to interpret from experimental data than those commonly used in current statistical analyses.