| Literature DB >> 28392451 |
Xiaohua Wan1, Tsvi Katchalski2, Christopher Churas2, Sreya Ghosh2, Sebastien Phan2, Albert Lawrence3, Yu Hao4, Ziying Zhou5, Ruijuan Chen6, Yu Chen4, Fa Zhang4, Mark H Ellisman7.
Abstract
Because of the significance of electron microscope tomography in the investigation of biological structure at nanometer scales, ongoing improvement efforts have been continuous over recent years. This is particularly true in the case of software developments. Nevertheless, verification of improvements delivered by new algorithms and software remains difficult. Current analysis tools do not provide adaptable and consistent methods for quality assessment. This is particularly true with images of biological samples, due to image complexity, variability, low contrast and noise. We report an electron tomography (ET) simulator with accurate ray optics modeling of image formation that includes curvilinear trajectories through the sample, warping of the sample and noise. As a demonstration of the utility of our approach, we have concentrated on providing verification of the class of reconstruction methods applicable to wide field images of stained plastic-embedded samples. Accordingly, we have also constructed digital phantoms derived from serial block face scanning electron microscope images. These phantoms are also easily modified to include alignment features to test alignment algorithms. The combination of more realistic phantoms with more faithful simulations facilitates objective comparison of acquisition parameters, alignment and reconstruction algorithms and their range of applicability. With proper phantoms, this approach can also be modified to include more complex optical models, including distance-dependent blurring and phase contrast functions, such as may occur in cryotomography.Entities:
Keywords: Alignment; Block face; Curvilinear; Electron; Inversion; Microscope; Reconstruction; Series; Simulator; Tilt; Tomography; Tracking; Warping
Mesh:
Year: 2017 PMID: 28392451 PMCID: PMC5773259 DOI: 10.1016/j.jsb.2017.04.002
Source DB: PubMed Journal: J Struct Biol ISSN: 1047-8477 Impact factor: 2.867