polyhedra, allowing averaging of the vertices to reduce noise and reveal substructure in a highly pleomorphic specimen [86]. Bharat et al. [83] modeled the
lattice of helical tubes of the HIV capsid protein, locating and averaging individual
unit cells to a resolution of 8 Å.
8.9 Conclusion
Ultimately we would like to know what reconstruction approach is the best to use.
The answer is complicated because it involves multiple issues that may be specific
to a particular situation. For instance, if the data originates from a non-FEG
microscope, it is likely resolution-limited to such an extent that CTF correction is
not feasible, and that most reconstruction algorithms will yield reasonable tomograms. However, if the aim is to do subtomogram averaging to high resolution on
data from an FEG microscope with a direct detector, then preprocessing (both real
space and CTF correction), and appropriate reconstruction (including denoising) are
required. What is appropriate depends to some extent on computational resources
and the targeted interpretation. If the goal is to generate tomograms fast or in large
numbers, the slow computation-intensive reconstruction algorithms are unsuitable.
Because denoising removes high frequency information, any form of it (including
iterative reconstruction techniques such as SIRT) may erode useful high resolution
details that are desired in subtomogram averaging. For these reasons, we try to
tailor the many different reconstruction methods to specific applications.
References
1. D.J. De Rosier, A. Klug, Reconstruction of three dimensional structures from electron
micrographs. Nature 217(5124), 130–134 (1968)
2. J. Radon, On the determination of functions from their integral values along certain manifolds.
IEEE Trans. Med. Imaging 5(4), 170–176 (1986). doi:10.1109/TMI.1986.4307775
3. J. Radon {Ü}ber die {B}estimmung von {F}unktionen durch ihre {I}ntegralwerte längs
gewisser {M}annigfaltigkeiten. Akad Wiss 69, 262–277. doi:citeulike-article-id:7680709
(1917)
4. G.T. Herman, Fundamentals of Computerized Tomography: Image Reconstruction from
Projections (Springer Publishing Company, Incorporated, 2009)
5. P.F.C. Gilbert, The reconstruction of a three-dimensional structure from projections and its
application to electron microscopy. II. Direct methods. Proc. R. Soc. Lond. B: Biol. Sci. 182
(1066), 89–102 (1972). doi:10.1098/rspb.1972.0068
6. R. Gordon, G.T. Herman, Reconstruction of pictures from their projections. Commun. ACM
14(12), 759–768 (1971). doi:10.1145/362919.362925
7. G.P. Vigers, R.A. Crowther, B.M. Pearse, Three-dimensional structure of clathrin cages in
ice. EMBO J. 5(3), 529–534 (1986)
8. R. Bracewell, Strip integration in radio astronomy. Aust. J. Phys. 9(2), 198–217 (1956). doi:
http://dx.doi.org/10.1071/PH560198
8 Tomographic Reconstruction from Electron Micrographs
231
lattice of helical tubes of the HIV capsid protein, locating and averaging individual
unit cells to a resolution of 8 Å.
8.9 Conclusion
Ultimately we would like to know what reconstruction approach is the best to use.
The answer is complicated because it involves multiple issues that may be specific
to a particular situation. For instance, if the data originates from a non-FEG
microscope, it is likely resolution-limited to such an extent that CTF correction is
not feasible, and that most reconstruction algorithms will yield reasonable tomograms. However, if the aim is to do subtomogram averaging to high resolution on
data from an FEG microscope with a direct detector, then preprocessing (both real
space and CTF correction), and appropriate reconstruction (including denoising) are
required. What is appropriate depends to some extent on computational resources
and the targeted interpretation. If the goal is to generate tomograms fast or in large
numbers, the slow computation-intensive reconstruction algorithms are unsuitable.
Because denoising removes high frequency information, any form of it (including
iterative reconstruction techniques such as SIRT) may erode useful high resolution
details that are desired in subtomogram averaging. For these reasons, we try to
tailor the many different reconstruction methods to specific applications.
References
1. D.J. De Rosier, A. Klug, Reconstruction of three dimensional structures from electron
micrographs. Nature 217(5124), 130–134 (1968)
2. J. Radon, On the determination of functions from their integral values along certain manifolds.
IEEE Trans. Med. Imaging 5(4), 170–176 (1986). doi:10.1109/TMI.1986.4307775
3. J. Radon {Ü}ber die {B}estimmung von {F}unktionen durch ihre {I}ntegralwerte längs
gewisser {M}annigfaltigkeiten. Akad Wiss 69, 262–277. doi:citeulike-article-id:7680709
(1917)
4. G.T. Herman, Fundamentals of Computerized Tomography: Image Reconstruction from
Projections (Springer Publishing Company, Incorporated, 2009)
5. P.F.C. Gilbert, The reconstruction of a three-dimensional structure from projections and its
application to electron microscopy. II. Direct methods. Proc. R. Soc. Lond. B: Biol. Sci. 182
(1066), 89–102 (1972). doi:10.1098/rspb.1972.0068
6. R. Gordon, G.T. Herman, Reconstruction of pictures from their projections. Commun. ACM
14(12), 759–768 (1971). doi:10.1145/362919.362925
7. G.P. Vigers, R.A. Crowther, B.M. Pearse, Three-dimensional structure of clathrin cages in
ice. EMBO J. 5(3), 529–534 (1986)
8. R. Bracewell, Strip integration in radio astronomy. Aust. J. Phys. 9(2), 198–217 (1956). doi:
http://dx.doi.org/10.1071/PH560198
8 Tomographic Reconstruction from Electron Micrographs
231