IDENTIFICATION OF SPHERICAL VIRUS-PARTICLES IN DIGITIZED IMAGES OF ENTIRE ELECTRON-MICROGRAPHS

New methods are described that should facilitate high-resolution (5-10 Angstrom) image reconstructions from low-dose, low-contrast electron micrographs of frozen-hydrated specimens and processing of large, digi tal images produced by new imaging devices and modern electron microsc opes. Existing techniques for automatic selection of images of individ ual biological macromolecules from electron micrographs are inefficien t or unreliable. We describe the Crosspoint method (CP), which produce s good quality solutions with relatively small miss rates and few fals e hits, and an extension of this method along with a procedure for ref ining its solution. Two algorithms for processing large images, one ba sed on image subsampling, the other on image decomposition, are descri bed. A large image is first compressed (e.g., by subsampling) and the CP method is applied to the compressed image to produce an initial sol ution. The information gathered at this stage is used to cut the origi nal image into subimages and then to refine the particle coordinates i n each subimage. An interactive environment for experimenting with par ticle identification methods is described. (C) 1997 Academic Press.