This protocol wraps e2refine2d.py EMAN2 program.
This program is used to produce reference-free class averages from a population of mixed, unaligned particle images. These averages can be used to generate initial models or assess the structural variability of the data. They are not normally themselves used as part of the single particle reconstruction refinement process, which uses the raw particles in a reference-based classification approach. However, with a good structure, projections of the final 3-D model should be consistent with the results of this reference-free analysis.
This program uses a fully automated iterative alignment/MSA approach. You should normally target a minimum of 10-20 particles per class-average, though more is fine.
Default parameters should give a good start, but are likely not optimal for any given system.
Note that it does have the –parallel option, but a few steps of the iterative process are not parallellised, so don’t be surprised if multiple cores are not always active.
Run the EMAN program to refine 2d.