Abstract: As theorists, we make predictions on the future evolution of a dynamical system based on known equations of motion.This is the traditional bottom-up approach. What if those equations of motion are unknown? Here we present a top-bottom approach to building dynamical models for complex systems such as proteins. The approach we present exploits a novel branch of Statistics -- called Bayesian nonparametrics (BNPs) -- first proposed in 1973 and now widely used in Data Science as the important conceptual advances of BNPs have become computationally feasible in the last decade. BNPs are new to the physical sciences. They make flexible (nonparametric) assumptions to efficiently learn models from complex data sets. Here we will show how BNPs can be adapted to address important questions in protein biophysics directly from data which is often limited by factors such as finite photon budgets as well as other fluorophore artifacts in addition to data collection artifacts (e.g. aliasing, drift). More specifically, we will show that BNPs hold promise by allowing complex spectroscopic time traces (e.g. smFRET, photon arrivals) or images (e.g. single particle tracking) to be analyzed and turned into principled models of protein motion -- from diffusion to conformational dynamics and beyond.
Biography: Steve Pressé was born in Montreal, Canada, and earned his Bachelor’s degree in chemistry from McGill (2000-2003). He first came to the United States to attend graduate school at MIT (2003-2008) where he worked on problems in the field of chemical physics with his advisor Prof. Robert J. Silbey. Steve pursued his postdoctoral studies at UCSF(2008-2012) under the guidance of Prof. Ken A. Dill. At UCSF, Steve broadened his interests from soft condensed matter into biophysics. Steve began as Assistant Professor at IUPUI (2013-2016) and moved to ASU as Associate Professor in 2017 joining the departments of Physics and the School of Molecular Sciences. Over the course of his collaboration with experimentalists, Steve began developing data analysis methods and developing models inspired by the tools of stochastic processes and statistical mechanics. His lab’s work is now focused on problems in Bayesian inference, frequentist data analysis methods, problems in protein dynamics and experiments on bacterial predator-prey communities. Steve has earned a number of fellowships, scholarships and awards over the course of his short career most recently earning the NSF CAREER in 2016 for his work on data analysis of imaging and spectroscopy methods.