CTF Refinement on symmetry expanded particles


I’ve been testing some things to improve my local refinements after symmetry expansion.

I have noticed that iterative refinements with smaller sampling angles/shifts improves the resolution a little bit. In addition, I’ve also tested running a round of local refinement on the symmetry expanded particles (using the mask from the local refinement) and also see a slight improvement in resolution. The map also qualitatively improves. I don’t do any job which requires global re-alignment of the particles.

I was wondering if it was considered OK to run local CTF refinement on symmetry expanded particles. From my point of view it makes sense, particularly in cases where the initial CTF estimation in the global refinement is poor (as in a very flexible region of a complex).

Any comment is more than welcome!

For local CTF refinement (i.e. defocus refinement) using the symmetry expanded particles could be beneficial, if the particle is large enough that different asymmetric units could have significant differences in defocus AND the AU themselves are large enough to fit the defocus accurately.

Otherwise, you don’t really want to do any kind of CTF refinement (or per-particle motion polishing) with a focused mask, it won’t work well compared to using the whole particle. Theoretically, the symmetry expansion shouldn’t lead to wrong values, but there would be no benefit.

PS CTF refinement is generally not warranted at resolutions worse than say 3.2 Å, unless there is something very wrong with the initial estimates.


Makes sense. Thanks for your answer!