Novel Step Detection Algorithms for Photobleaching Analysis of Protein Complexes with Many Subunits

Abstract

Single molecule photobleaching of fluorescently labeled protein complexes is an effective technique for counting constituent subunits and has been applied to determine the stoichiometry and oligomerization of several different transmembrane proteins including voltage- and ligand-gated ion channels, as well as the makeup of the cellulose synthesis complex. Fluorophore bleaching occurs as a random process, resulting in discrete intensity drops over time. Step detection algorithms can be used to identify these steps within noisy signals in order to determine the total number of fluorescent subunits within a protein complex. Reliable identification of steps can become difficult however, when high numbers of subunits are present. Here, we present two step detection algorithms, one based on modified t-testing and another based on the Bayesian Information Criterion (BIC), that perform at higher levels of precision than existing algorithms and account for temporal changes in variance within a signal, which is expected in photobleaching experiments. We also present a method for determining a unitary step amplitude associated with bleaching of a single fluorophore based on a Gaussian Mixture Model.