17. Local Deblur Sharpening protocol

Protocol designed to apply LocalDeblur, the automatic local resolution-based method that increases map signal at medium/high resolution [Ramírez-Aportela et al., 2018], in Scipion. Unlike similar approaches, LocalDeblur does not need any prior atomic model, avoiding artificial structure factor corrections. Since the map gets much more interpretability, the modeling process results much easier. This type of sharpening is recommended for maps showing a broad range of resolutions, as it is usually common with membrane proteins or macromolecules highly flexible. In all those cases, applying a global sharpening method, like Relion postprocess, does not optimize the result when you have very different local resolution values because a global operation cannot improve all parts of the map. However, when changes in resolution are small there isn’t almost difference between applying a global or a local sharpening method. For example, in maps with high symmetry, like viruses, resolution is quite homogenous. The absence of high differences in resolution determines that at high resolution the results of global or local sharpening methods are almost the same.

  • Requirements to run this protocol and visualize results:
    • Scipion plugin: scipion-em
    • Scipion plugin: scipion-em-xmipp
    • Scipion plugin: scipion-em-chimera
  • Scipion menu:
    Model building -> Preprocess map (Fig. 17.1 (A))
    Protocol **xmipp3-localdeblur sharpening**. A: Protocol location in *Scipion* menu. B: Protocol form.

    Fig. 17.1 Protocol xmipp3-localdeblur sharpening. A: Protocol location in Scipion menu. B: Protocol form.

  • Protocol form parameters (Fig. 17.1 (B)):
    • Input map: Unfiltered electron density map previously downloaded or generated in Scipion.
    • Resolution Map: Resolution map generated by protocols like xmipp3-local MonoRes. The resolution value in the corresponding voxel of the Input map is assigned to each voxel of the Resolution Map.
    • lambda: Since LocalDeblur is based on an iterative formula repeated until a convergence criterion is reached, \lambda is the step size advanced parameter that modulates the speed of convergence. The default value, \lambda = 1, indicates that the method itself establishes automatically the value of \lambda. Although the default value is small enough to guarantee the convergence and large enough to speed it up, \lambda value can be increased by the user to accelerate the convergence process. Unlike the default value, that grows along the convergence process, \lambda value selected by the user will be maintained constant. Falling into a local minimum is a risk derived of increasing the convergence speed.
    • K: Weight assigned to the difference between the local resolution and the spatial frequency of the center of each bandpass filter. This difference weighted by K is the base to compute the local weight of each channel in the filter bank, that correlates the input map with the sharpened map. The bigger the value of K, the lower the weight of each channel in the filter bank. Maximum weights are obtained when local resolution and spatial frequency of the center of each bandpass filter show identical values. As it has been empirically observed K=0.025 produces good results for most of tests performed. No big differences have been detected with K values ranging between 0.01 and 0.05. In the particular case of low resolution maps (lower than 6 Å), 0.01 seems to be a good choice.
  • Protocol execution:
    Adding specific map/structure label is recommended in Run name section, at the form top. To add the label, open the protocol form, press the pencil symbol at the right side of Run name box, complete the label in the new opened window, press OK and, finally, close the protocol. This label will be shown in the output summary content (see below). If you want to run again this protocol, do not forget to set to Restart the Run mode.
    Press the Execute red button at the form bottom.
  • Visualization of protocol results:

    To visualize the total number of sharpened maps generated according to the number of iterations until convergence, press with the right mouse the black arrow placed in the lower part of the Scipion framework (Protocol output: xmipp3 - localdeblur sharpening -> ouputVolumes). The option Open with DataViewer will appear, select it. The sharpened map epochs will be detailed. The sharpening algorithm stops when the difference between two successive iterations is lower than 1%, thus generating a variable number of maps before stopping. You can select any of them and visualize the slices with ShowJ, the default Scipion viewer. To visualize the slices press the symbol that appears below File in the main menu. The ShowJ window menu (File -> Open with Chimera) allows to open the selected map in ChimeraX graphics window.
    Nevertheless, ChimeraX graphics window can also be opened to compare the input and the last iteration sharpened map. After executing the protocol, press Analyze Results and ChimeraX graphics window will be opened. Volumes are referred to the origin of coordinates in ChimeraX. To show the relative position of the volumes, the three coordinate axes are represented; X axis (red), Y axis (yellow), and Z axis (blue) (Fig. 5.3). Coordinate axes, the imported volume and the last sharpened one (sharpenedMap_last.mrc) are model numbers #1, #2 and #3, respectively, in ChimeraX Models panel. Volume coordinates and pixel size can be checked in ChimeraX main menu Tools -> Volume Data -> Map coordinates: Origin index/ Voxel size.
  • Summary content:
    • Protocol output (below Scipion framework):
      xmipp3 - localdeblur sharpening -> ouputVolumes;
      SetOfVolumes (number of items, x, y, and z dimensions, sampling rate).
    • SUMMARY box:
      LocalDeblur Map.