RotationalSpectraClassification

Rotational Spectra Classification

• 2D-alignment for all particles in selection file images.sel.
• Find the centre-of-symmetry in the average of the aligned images.
• Calculate the rotational spectra for all individual particles.
• Calculate a self-organizing map of all rotational spectra.

CRITICAL STEP: The self-organizing algorithm proceeds from an initially high value of the regularization parameter (-reg0) to a lower value (-reg1) in a user-defined number of steps (-steps). Too high regularization values result in too smooth output maps that do not explain the variance in the data, while too low values yield maps that are not organized. Typically, one repeats this calculation multiple times with varying annealing parameters, in order to optimize the output map.

• Inspect the self-organizing map, and identify distinct classes.

Example Data

• Input
  • Gallery of images.

File rotational_spectra_classification.tar.gz contains the images obtained after preprocessing and manual picking of 18 tilted pairs. The script with suitable parameters for the test data is xmipp_protocol_rotspectra.py. simC3.sel and simC6.sel are the selection files for the 3-fold and 6-fold classes respectively.

• Output
  • One selection file for each selected group
  • Auxiliary files used by the analyze results.

File output_rotational_spectra_classification.tar.gz contains the results of executing therotational_spectra script.

Pressanalyze results button for visualizing the rotational spectra classification.

Gallery of aligned images (top right) and self organizing map of the rotational spectra (bottom)

NOTE1: since this protocol uses random seeds, the results may vary between different runs.

NOTE2: Better results are obtained if each class is realigned and re-classified (the results provided here have gone through two of this cycles).

USER's COMMENTS