UC Berkeley

Following the popularization of super-resolution microscopy for biological research, the advent of functional super-resolution microscopy enabled decoding high-throughput single-molecule signals in single-molecule localization microscopy (SMLM) and extracting multiple physical parameters beyond the structure of the target system. Among those parameters, the diffusivity of single molecules has provided deeper insight into biological systems as diffusion is one of the key mechanisms that carry out life processes and as the diffusivity reflects multiple physicochemical parameters and thus properties of the microenvironment. This dissertation describes applications of the recently developed wide-field ultrahigh-throughput diffusion coefficient analysis and a new method that enables the equivalent measurement without sophisticated hardware setup. By extensively tracking diffusion behaviors of proteins in nanometer-sized matrix structured hydrogel, I drew a comprehensive picture of how different sizes of meshwork affect the protein diffusion coefficients, as well as how such an effect is differentiated by the relative size of proteins to the meshwork. Then, I introduce a new approach to extract diffusion coefficients out of conventional SMLM by incorporating single-molecule image processing through a convolutional neural network. In the last chapter, I summarized efforts on taking the SMLM diffusion analysis to a greater extent.