Skip to content

Tomogram Reconstruction using NovaCTF

Final tomogram reconstruction for subtomogram averaging will be performed using novaCTF, which allows for 3D CTF-correction during reconstruction. TOMOMAN will generate the appropriate scripts and output directories for running novaCTF at the desired binnings.

For this tutorial, we will walk through setting up and running novaCTF. We will directly reconstruct an 8x binned tomogram, and if time permits, we can later reconstruct a 4x binned tomogram for higher-resolution averaging.

Run NovaCTF

Open tomoman_novactf.param.

  1. The directory parameters should already be correct.

  2. The parallelization parameters determine how jobs are split between cores. Set n_cores to 16.

  3. Stack parameters are parameters for generating the aligned stacks prior to tomogram reconstruction.

    • ali_dim allows for resizing, but we recommend using the full image size: 3712,3712.

    • erase_radius is for gold fiducial erasing. Since we used autmoated tilt-series alignment, we don't have a model for the gold fiducials. Set this to none.

    • taper_pixels is used to taper the edges of the rotated images when generating an aligned stack; 100 is usually sufficient.

    • ali_stack_bin is used to bin the image stack before reconstruction. If you want tomograms of several binning factors, it is possible to reconstruct at the lowest binning, then bin the tomograms. While this produces the best results, it is often more computationally efficient to bin the stack to the target binning for the tomogram, and perform this multiple times if necessary.

    For this step, set ali_stack_bin to 8.

  4. The 3D CTF correction parameters set how novaCTF will perform 3D CTF correction. We always recommend using the dose-filtered stack (process_stack = dose-filtered) and correcting CTF using phase flipping (correction_type = phaseflip). For defocus_step, smaller steps produce more precise results at the cost of more computation time (see the novaCTF publication for more information). For this tutorial, set it to 50.

  5. Tomogram reconstruction parameters have some specifics on how to perform the reconstruction. We typically skip radial filtering. The tomo_bin parameter allows you to set the final binning factors desired. For this tutorial, set this to tomo_bin = 8.

    NOTE: The minimum allowed value for binning is equal to the ali_stack_bin. E.g., if ali_stack_bin is set to 4, the minimum allowed value here is 4.

  6. The output_dir_prefix sets the name of the tomogram output directories, which will be placed within the root_dir. For instance, bin 8 tomograms will be placed in: [root_dir]/[output_dir_prefix]_bin8/. For this tutorial, set this to novactf_.

  7. The additional parameters include the subset_list, which allows for reconstructing a subset of tomograms. You can leave it as none to reconstruct all non-skipped tomograms in the tomolist.

  8. Fourier3D is a program for Fourier cropping volumes written by Beata Turoňová. The f3d_memlimit parameter sets a limit to how much memory Fourier3D can use; more memory allows for faster computation times. Since we will not be binning tomograms, this parameter is ignored.

  9. NovaCTF’s approach to CTF-correction assumes that the center of mass is at the center of the tomograms. If this is off, the reconstructed tomogram will contain a systematic error in all planes. To refine the tomogram center, novaCTF allows you to generate an offset value for recentering. TOMOMAN can take an input STOPGAP motivelist, and use the center of mass of the particles from each tomogram as the refined center. Since we have no such motivelist now, this can be left off.

  10. Run novaCTF in the TOMOMAN console.

    tomoman(pwd,'tomoman_novactf.param');

  11. After running novaCTF or copying the precomputed tomograms, you can preview your them in 3dmod, for example:

    3dmod novactf_bin8/TS_01_dose-filt.rec

Preprocessing in TOMOMAN is now finished!