Powered by Indico
v3.2.3
In no area of science emergent complex behavior is more ubiquitous than in biology. Traditional approaches from physics encounter difficulties in describing such processes, which are in nature nonlinear, stochastic, dissipative and without detailed balance. In this talk, I will focus on two intriguing but distinct questions: How precise behaviors of single cells are emergent from noisy intracellular chemical reactions, and how complex diseases are emergent from underlying molecular interactions.
For the first question, we derived fundamental limits for timing in chemical reaction systems, by greatly generalize an elegant proof for bounds on first passage time regarded as “the most important result” in the field and took mathematicians more than 30 years to find. We then turn to practical reaction mechanisms and show that an exceedingly simple mechanism in fact can get close to the optimal precision. A recently developed high-throughput, long-term, single cell time-lapse imaging and screening platform enables for the first time detailed characterization and isolation of dynamic phenotypes of single cells. Experimental results show that the simple but near-optimal mechanism is in fact used in all three systems that have been studied in depth using such a platform.
For the second question, we developed a novel SDE decomposition framework, where a dual role potential function, serving both Lyapunov function for deterministic counterpart dynamics and energy function in Boltzmann-Gibbs type steady state distribution, can be obtained for general stochastic dynamics. We explicitly constructed such potential functions for limit cycles and chaotic attractors, which was deemed impossible by many famous mathematicians. The framework enables analysis of large-scale network models for complex diseases such as cancer. We built core endogenous network models for prostate cancer, acute promyelocytic leukemia and colorectal cancer. By examining the nonlinear stochastic network dynamics, we found that robust states corresponding to normal physiological and abnormal pathological phenotypes, including cancer, emerge naturally. The network models recapitulate known clinical observations and can help us search for new therapies on cancer, e.g. drug combinations. Our studies indicate that stochastic dynamics may lay at core in rationalizing biological complexity.
Dr. Yuan received dual undergrad degrees in physics and computer science (ACM honors class, Zhiyuan College) from SJTU. He completed his PhD under supervision of Dr. Ping Ao in 2016, during which he was trained as a theorist in both areas of physics and biology. He then joined Dr. Johan Paulsson’s lab at Harvard Medical School as a postdoc and started doing single cell experiments. He is currently working with Dr. Adam P. Arkin at University of California, Berkeley. Dr. Yuan’s research interest includes understanding genesis and development of complex diseases, deciphering precise control mechanisms in single cells, developing theoretical tools for stochastic dynamics, as well as single cell time-lapse imaging and screening.