Carrington, Walter 9724611 Project Summary: A plan is presented to develop hardware, algorithms and software to enhance and extend the capabilities of the digital imaging microscope (DIM), an approach to 3-D fluorescence microscopy which combines wide field microscopy with high speed CCD cameras, fast focus change, deconvolution algorithms and visualization and analysis software. The plan has been developed in response to the limitations imposed on biological experiments by even the most advanced current systems. Enhancements to an existing ultra-fast 3-D microscope will be implemented, including increasing the number of high speed images that may be obtained before slow readout is necessary. New computational methods for 3-D image restoration(deconvolution) will be developed and tested that use L^p smoothing with non-negativity to achieve a lateral resolution of 75nm and axial resolution of 250nm. New software will be written to automate the calculation of a sub-sampled point spread function used in this super- resolution image restoration approach. We will develop a new parallel implemention of these algorithms on a single 4- processor Silicon Graphics, Inc. (SGI) Origin 200 computer using SGI's software development tools. Visualization tools that increase 3-D rendering speeds by 100x using hardware texture mapping on low end SGI 02 workstations will be developed. Finally, new methods for analyzing the statistical significance of spatial and temporal patterns in images will be developed. A number of biological projects will be carried out on this new instrument addressing key problems which are not soluble by currently existing instruments. For example, the pathway through the cell that glut4 glucose transporters take when they are sent to the plasma membrane in response to insulin will be directly visualized and the dynamics of assembly of centrosomal proteins and consequences of assembly and disassembly during the cell cycle for centros ome function will be determined. Biologists and engineers will be trained in optical methods and digital imaging microscopy through involvement in biological projects and instrument development respectively. The work we propose will enhance the sensitivity, resolution, numerical accuracy and visualization and analysis capabilities of the digital imaging fluorescence microscope. It will thus extend the range of questions that biologists can answer with it.
