Bayesian Nonparametrics for single molecule Biophysics

Cells contain 1 molecule of DNA, and sometimes as few as tens of RNAs and proteins of various species. In this regime, we need ways of describing events involving individual molecules. For this reason, we will start by describing the field of single molecule biophysics and how single molecule approaches have allowed us to glean insight free of bulk averaging inherent to classical methods. We will also show that the analysis of single molecule data often presents a model selection problem (where many competing models can explain the data). We will discuss both experimental and theoretical approaches to resolving model selection. In particular, we will focus on Bayesian nonparametrics, and in particular, Dirichlet and Beta-Bernoulli processes, to address the model selection problem and show that it can be both computationally efficient and suggest alternate strategies by which data may be collected and analyzed. If time allows, we will discuss strategies to track multiple particles in crowded environments using Bayesian nonparametrics.

Bio
https://labpresse.com/