Fluorescence microscopy is one of the major tools to study the function of proteins in a cellular environment. Recent advances in the fluorescent labeling of proteins, in imaging detectors and hardware control have made it possible to image individual protein molecules in a live-cell environment. Results from the imaging of the exocytosis of single antibody molecules from endothelial cells will be presented.
Single molecule experiments pose significant technical problems both from the biological and the engineering point of view. At the technical level, the extremely low signal to noise level accounts for many of the problems that are encountered during the analysis of single molecule microscopy data. A fundamental question is how accurately the location of a single molecule can be determined using a fluorescence microscope. By posing this problem as one of parameter stimation for a marked counting process, we have derived performance limits for a fluorescence microscope. These performance limits give insights into appropriate choices for experimental design.
Deconvolution plays an important role in the analysis of fluorescence microscopy image data to remove blur introduced by the optics of the microscope. In the use of deconvolution algorithms, experimentally acquired point spread functions are often preferred over theoretically derived ones due to their higher accuracy. The implementation will be discussed of a subspace based algorithm to smooth experimental point spread functions. The reconstructed images with a smoothed point spread function are significantly improved over images that are deconvolved with a non-smoothed point spread function.
If time permits, signals and systems approaches will be discussed for the analysis of molecular interactions using biosensors.
Back to Control Seminars Page.