Single molecule microscopy is a relatively novel technique that allows individual molecules to be imaged using the optical microscope. It is of major interest to scientists because it enables the observation of cellular processes at the molecular level, a task that is not achievable with classical optical imaging approaches due to the cellular processes being obscured by averaging and bulk effects. Single molecule tracking and localization-based superresolution microscopy (PALM/STORM) are two of the main examples of single molecule imaging experiments. The improvement of these techniques over classical imaging approaches is directly dependent on the accuracy with which the position of the single molecules in the sample can be estimated. This leads to two major challenges. One, the design of experiments such that the best possible data can be obtained. Two, the analysis of the data in the most optimal fashion. The technical difficulties of single molecule experiments are to a large extent due to the very low photon count in the presence of significant noise sources such as scattered photons and camera readout noise. The proposed project aims to implement algorithms to permit estimation tasks such as the localization of a molecule to be carried out with the best possible accuracy. An important part of the project is a Java-based software framework for the analysis of single molecule experimental data. At its core will be a carefully designed software architecture for the analysis algorithms that will allow in a streamlined fashion the incorporation of different models for the image profiles of single molecules, different camera-dependent data generation processes, different data pre-processing steps, and different estimation algorithms. By employing graphical processing units, advantage will be taken of the massively parallel structure of the overall data analysis. Exploiting our earlier information-theoretic results, we will also develop approaches and supporting software to allow the microscopist to analyze the impact of parameters such as exposure time, excitation level, magnification, and pixel size, and by doing so, design an experiment that will produce data that is optimized for the subsequent analysis. This will be complemented by methods to analyze the quality of the objective lens, the performance of dichroic filters, the noise level of cameras, etc.
The specific aims are Aim 1: to develop a single molecule data analysis platform including modules for single molecule tracking and single molecule superresolution microscopy, and Aim 2: to develop a framework for experimental design and instrument characterization in single molecule microscopy. The proposal originates from a spin-out company of Texas A&M University that seeks to further develop and commercialize single molecule methodologies developed over more than a decade.
Single molecule microscopy can reveal major new insights into biological processes at the molecular level in living cells, such as the interactions of anticancer drugs with cancer cells. However, this new method of microscopy requires complex analysis of the acquired image data to produce the desired results. The current proposal aims to provide new methods such as algorithms and software to accomplish this task.