CFE 2019: Start Registration
View Submission - CMStatistics
B1214
Title: Doubly stochastic point processes in time and space as a model for photoactivated localization microscopy data Authors:  Ute Hahn - University of Aarhus (Denmark)
Louis Gammelgaard Jensen - Aarhus University (Denmark) [presenting]
Abstract: Photoactivated localization microscopy (PALM) is a revolutionary technique that promises to localize single molecules, and increases resolution from 300nm to 20nm. The method uses photoactivatable fluorophores attached to the molecules of interest. An image is obtained from fluorescence photons that are registered in a video. Only few molecules are activated by a laser beam at a time, thus only few photons are registered in each video frame. This makes it possible to distinguish between photons coming from different molecules. Reconstruction of single molecule positions in PALM is crucially based on the assumption that the fluorophores are activated independently of each other. It is known though, that several photons can be emitted from the same molecule over time, an effect called photoblinking. This may lead to artifacts in the interpretation of the raw data. We set up a doubly stochastic model for the data, with a latent spatial point process representing the molecules. This process serves as parent process for a space-time cluster process, reflecting photoblinking. We present statistical methods for estimating the model parameters, allowing us to separate contributions from molecules and blinking artifacts.