Hi,
I have a query about how cryosparc deals with anisotropic magnification.
Anisomag can manifest as an apparent astigmatism, and therefore (apart from other effects) can lead to erroneous astigmatism estimates during the initial round of CTF estimation (Description and comparison of algorithms for correcting anisotropic magnification in cryo-EM images - ScienceDirect).
Does CS alter the per-micrograph astigmatism values to account for this effect? And if an anisomag matrix has been refined, would it be possible to provide this to Patch CTF to obtain improved CTF estimates in cases where anisomag is a significant factor?
Cheers
Oli
Great question @olibclarke,
Once estimated in Global CTF Refinement, anisotropic magnification is used to correct for the effect of mis-estimated astigmatism at the particle level. If anisomag estimation is turned on in Global CTF Refine, then the defoci & astigmatism of each particle (in particle.ctf) will be updated to values they would have as if they have been estimated with the astigmatism known from the beginning, assuming that the anisomag is not so large as to cause the initial astigmatism estimation to fail completely (e.g. by having the optimal value be out of the range that CryoSPARC searches over in patch CTF). If that makes sense 
However, micrograph CTF data (exposure.ctf) never sees the anisotropic magnification, nor does it see the updated defoci after anisomag is known. Micrograph CTF data also never get updated with any of the high order aberrations. It may be possible that accounting for very significant anisomag during patch CTF could improve CTF estimates slightly, but we don’t yet have a strong reason to believe this would make a significant effect. Do you have an example where there is significant anisotropic mag and you suspect Patch CTF could perform significantly better after accounting for this?
Best,
Michael
1 Like
Hi Michael - no this part:
If anisomag estimation is turned on in Global CTF Refine, then the defoci & astigmatism of each particle (in particle.ctf) will be updated to values they would have as if they have been estimated with the astigmatism known from the beginning, assuming that the anisomag is not so large as to cause the initial astigmatism estimation to fail completely (e.g. by having the optimal value be out of the range that CryoSPARC searches over in patch CTF). If that makes sense
Is what I wasn’t clear on, and now makes perfect sense, thanks! No, I don’t think there would be any need to update the micrograph CTF, given this is taken care of at the particle level 
1 Like