Help with Symmetry Alignment in non-uniform Refinement (C2 Symmetry Misaligned)

3D Refinement
1

Hi,

I’m currently analyzing a protein complex, which is based on biochemical reasoning and supported by 2D classification, expected to have C2 symmetry.

My ab initio reconstruction (with C1 symmetry) yielded me a map that shows just over half of the complex. When I run non-uniform refinement:

  1. Using C1 symmetry, I get a reasonable map at ~5.5 Å resolution with identifiable domains.
  2. But using C2 symmetry, the resulting map looks completely wrong, the symmetry axis appears to be misidentified or incorrectly aligned.

I’m unsure how CryoSPARC determines the symmetry axis in such cases. Is there a way to manually specify or guide CryoSPARC to the correct symmetry axis during refinement?

One additional detail: my dataset does exhibits some degree of orientation bias, and I wonder if this might be contributing to the problem possibly leading CryoSPARC to an incorrect symmetry alignment.

Any suggestions on how to deal with this, or tips for troubleshooting incorrect symmetry alignment in CryoSPARC, would be greatly appreciated.

While this is not a dealbreaker, I would like to learn how to deal with issues like this. especially if it can help improve the resolution.

Thanks in advance!

2

Hi @skumarb1
Well, please post 2d pics, abinitio and NU-refine volume so that the community can guide you whats going wrong.
Are you sure that your protein has C2 symmetry and not pseudo symmetry? In the later case there are various ways to address them. You can apply C2 in NUrefine by unchecking “align symmetry” and see if it improves.
What is the pixel size you collected? 5.5 is faraway and means to me that there is still junk to be removed. You may perform 3D classification followed by NUR and see if it improves.
Sometimes bypassing 2D and using heterogeneous classification can lead to improved orientation.

3

The in-job symmetry alignment can get it quite wrong. If the map looks off-centre, try symmetry alignment using Volume Alignment Tools or download the C1 map, align with RELION and re-import to use as a reference, then turn off symmetry alignment in the job (don’t turn off symmetry, just alignment!)

Symmetry usually will help mitigate orientation bias, provided it isn’t too extreme. The angular assignment plot would be useful as well as the things Das suggests.

1 Like