Preprocess images

RELION will work best if your data set is:

• Of high signal-to-noise ratio (take great care in sample preparation and data collection)
• Clean from false particles (no images are discarded during refinement).
  • Xmipp implements a utility called xmipp_sort_by_statistics to sort particles which comes handy in the cleaning of a data set.
• Unmasked (masking is performed internally)
• Non-interpolated (prevent any prior rotations/translations: use the originally scanned pixel values)
  • If downscaling is necessary because of memory issues: use a window-operation in Fourier-space, not a convolution (e.g. with rectangle/B-spline).
  • Xmipp implements the Fourier-space downscaling in the xmipp_scale program with the -fourier option.
• uncorrected for CTF (this is done internally)
  • If your data have previously been phase-flipped, that's OK: just set the corresponding option
  • If your data have previously been pre-Wiener filtered, that's a VERY bad thing to do in general: go back to the original data.

RELION does not discard any image during refinement. Therefore, data sets should be clean from false particles. Masking and CTF correction are done internally, so in the preprocessing images do not need to be masked, nor CTF phase-flipped or Wiener-filtered. The images should be normalised, but the exact procedures do not matter much, as also normalisation errors may be corrected inside the refinement.

CTF phase-flipping