Particle Alignment¶

Regardless of the method eventually utilized for 3D reconstruction, a good starting point for single particle EM investigations is 2D alignment and classification of the dataset. This type of analysis intimately aquants the scientist with the types of particles, distribution of views, and the relative amount of "junk" contained in the dataset.

Many methods exist for particle alignment and classification, but the general workflow for this type of analysis is as follows:

1. Align particles using reference-free or reference-based approaches.
2. Obtain a mathematical descriptor of the variance within the dataset using a feature analysis algorithm.
3. Sort the dataset into classes according to an interimage variance metric using a clustering algorithm.

Please see the links for each of these three steps for more information on the various packages available for performing each step in the pipeline.