Validated Near-Atomic Resolution Structure of Bacteriophage Epsilon15 Derived from Cryo-EM and Modeling
ELECTRON MICROSCOPY
Sample
Bacteriophage epsilon15
Specimen Preparation
Sample Aggregation State
PARTICLE
Vitrification Instrument
FEI VITROBOT MARK II
Cryogen Name
ETHANE
Sample Vitrification Details
Blot before plunging into liquid ethane (FEI VITROBOT MARK II).
3D Reconstruction
Reconstruction Method
SINGLE PARTICLE
Number of Particles
14000
Reported Resolution (Å)
4.5
Resolution Method
FSC 0.143 CUT-OFF
Other Details
The gold standard definition for the resolution estimate was adopted whereby the particle images were split into two subsets at the onset of image pro ...
The gold standard definition for the resolution estimate was adopted whereby the particle images were split into two subsets at the onset of image processing and the datasets were individually reconstructed and then combined after determination of the resolution estimate. Independent initial models were built de novo and used for the subsequent particle refinements in each of the two subsets of particle images. The Fourier Shell Correlation (FSC) between the two independently determined reconstructions was computed and indicated a resolution of 4.5 Angstrom using the 0.143 threshold for the combined dataset. (Single particle details: Individual particles (720x720 pixels) were first automatically selected using the ethan method followed by manual screening using the EMAN boxer program. A total of 54161 particles were selected for initial processing. The selected particles within a micrograph were incoherently averaged to generate 2D power spectra for contrast transfer function (CTF) parameter determination. CTF parameters were first automatically estimated and then visually verified using the EMAN1 ctfit program. Defocus values range from 0.5 to 2.5 um. The data set was divided into two data subsets for the following reconstruction steps. The particle images were first binned 4x for initial model building and initial determination of orientation and center parameters. The initial model was built de novo by iterative refinement of a subset of 300 particles randomly selected from the half data set with randomly assigned initial orientations. The initial orientations of all particles in each of the half data sets were determined using the EMAN1 projection matching program classesbymra with an angular projection step size of 3 degrees. The orientations were then refined to higher accuracy using the program jalign, which is based on simplex optimization of matching between the particle image and model projections. The particle orientation parameters were then transferred to particles binned at 2x and ultimately to particles without binning for further refinements. In the last stage of refinement, magnification, astigmatism, and defocus parameters were also included. 3D maps with icosahedral symmetry enforcement were reconstructed using a newly developed program, j3dr, using the EMAN2 library and parallelized with message passing interface (MPI) to speed up the reconstruction process. These steps were iterated until the refinement converged. The map for each data subset was reconstructed from ~7000 particles by removing particles with poor alignment scores and unstable alignment parameters. The resolution of the map was evaluated using the Fourier Shell Correlation (FSC). Only the icosahedral shell region was included in this FSC analysis by masking out the external background noises and the internal DNA densities using soft masks with a half width of 6 Angstrom. The final map of the entire dataset was then built from ~14000 particles by combining these two subsets of particles.) (Single particle--Applied symmetry: I)
