Abstract
The tomographic reconstruction from whole-cell electron tomography, which is used in the study of viruses in situ, is generally noisy and geometrically distorted due to low electron dose and incomplete projection data. Identical copies of virus particles are expected to be present in a cell, thus by combining data from different virus particles at the same lifecycle stage, a higher resolution structure can be obtained. This classification and data fusion problem is solved using a maximum likelihood estimation approach via an expectation-maximization algorithm. The classification result agrees with the visual impression from the tomographic data and the final reconstruction of each class agrees with the reference structure from single particle cryo-EM analysis.
© 2009 Optical Society of America
PDF ArticleMore Like This
V. L. Lo and R. P. Millane
STuB2 Signal Recovery and Synthesis (SRS) 2009
Frank Vollmer and Stephen Arnold
LSMG2 Laser Science (LS) 2009
Xiaowei Zhuang
SF2 Biomedical Topical Meeting (BIOMED) 2006