|  | |
| Title: | Molecular Architecture of the ATP-Dependent Chromatin Remodeling Complex SWR1 by 3 Dimensional Electron Microscopy |
| Authors: | Nguyen VQ, Ranjan A, Stengel F, Wei D, Aebersold R, Wu C, Leschziner AE |
| Sample: | ATP-dependent chromatin remodeling complex SWR1 |
| Method: | Single particle reconstruction (28 angstroms resolution) |
| Related entries in frame: | EMD-5638 |
|
| |
| |
| Processing |
| Protocol: | Orthogonal Tilt Reconstruction |
| Software: | EMAN1, EMAN2, IMAGIC, SPIDER |
| CTF correction: | CTFTILT (Grigorieff) for initial model; EMAN2 for projection-matching refinement |
| Number of particles: | 18000 |
| Imposed symmetry: | C1 |
| Resolution by author: | 28 Å
|
| Resolution method: | FSC at 0.5 cut-off |
| Processing details: | In order to extract the molecular images from the micrographs, particles were windowed out in one set of micrographs (-45 degrees) using the Boxer interface in EMAN1. Custom-built SPIDER scripts were used to calculate alignment parameters between the +45 degree and -45 degree micrographs and to extract the tilt mates in the +45 degree micrographs. The Contrast Transfer Function (CTF) was estimated and corrected for using the program CTFTILT and the SPIDER command TF CT. Single particles were binned by 2, resulting in a pixel size of 3.3 Angstrom. The +45 degree and -45 degree datasets were combined into a stack of ~32,000 particles and 2D alignment and classification were performed in IMAGIC. Initial models were computed from classes containing 100-200 members using the Orthogonal Tilt Reconstruction approach as described.
For projection-matching refinement, the OTR models were initially refined against 2D class averages of cryo-negative data. To generate the class averages, particles were extracted from the micrographs as described above and CTF estimation and phase flipping were performed using the EMAN2 workflow. The particles were then binned by 2, resulting in a pixel size of 3.45 Angstrom. Approximately 32,000 particles were subjected to reference-free 2D alignment and classification in IMAGIC. In order to minimize heterogeneity, classes were generated with relatively few (15-20) particles.
The OTR models were filtered to 80 Angstrom resolution and 15-23 iterations of projection matching refinement were performed using angular steps of 250, 200, 150, 100, and 80-50 against 2D class averages in SPIDER using the AP SH and BP 32F commands. To minimize noise in the reconstructions, a threshold mask calculated for 500% to 150%of the theoretical molecular weight of the sample (1.0 MDa) was applied. The mask was gradually tightened throughout refinement and its filtration was determined by the resolution of the 3D map, computed according to the 0.5 FSC criterion. Refinement results were stable after 15 iterations, and the resolutions of the 3D maps were 50-60 Angstrom. The resulting 3D maps (without additional filtration) were then similarly refined against single cryo-negative particles. For this step, 15 iterations of projection-matching refinement were carried out at angular steps of 250, 210, 180, 150, 130, 110 and 100-40. Threshold masks computed for 500% to 100% of the MW were also utilized. |
| Scanned images: |
|
| Fitting: |
|
|
|
| | |
|
