Improving the Scope and Quality of Single-Molecule Data Analysis

[摘要] In this dissertation, I extend the scope and quality of the information that may be gained from a single-molecule imaging data set. First, I investigate a more analytically accurate point spread fitting function for use in determining not only the location of fluorescent molecules but also its in-frame displacement. This contrasts with previous work that determined the displacement through an empirically derived calibration function. Some improvements are shown to be possible by using the new sum of error functions (SErf) fitting function instead of the previously used asymmetric Gaussian. Secondly, I analyze the extremely quickly diffusing molecules inside bacteria. These diffusing molecules are found to be so blurred by in-frame motion so as to prohibit accurate localization and thus preclude the subsequent use of single particle tracking for determining the molecules’ diffusive properties. Instead I implement for the first time in bacteria spatio-temporal image correlation spectroscopy (STICS), which does not suffer from in-frame blur in the same way as SPT. I did, however, find two biases inherent to STICS and describe their causes to be experimental deviation from the expected diffusive step size distributions due to in-frame motion and tight confinement. Finally I consider the multi-step fitting process commonly used to determine the mean squared displacement (MSD) of diffusing molecules via the cumulative probability distribution (CPD). I find that by combining the CPD and MSD fits into a single, multi-domain fit, the number of free parameters in the fitting procedure can be reduced from dozens to fewer than 10. This reduction in the fitting degrees of freedom is due to combining redundant parameters, and both improved the precision of diffusion coefficient estimation as well as the likelihood that the fitting procedure produced physically relevant results. These three areas of focus expand the scope of single-molecule super-resolution data analysis by simplifying a method of determining instantaneous displacements of point light sources, and improve the quality of both spatiotemporal image correlation spectroscopy in bacteria and single-particle tracking analysis via cumulative probability distributions of squared step sizes.

[发布机构] University of Michigan

[效力级别] Physics

[关键词] Single-molecule super-resolution microscopy analysis methods;Physics;Science;Biophysics