Pseudosymmetry in a D3 complex. can symmetry relaxation help?

3D Refinement
1

Hi everyone,
I am dealing with a D3 (hexameric) complex. After doing some preliminary refinements in C1, I noticed that one subunit out of six is definitely poorly defined compared to the others. I am wondering what might be the best workflow in this case.
My idea is: 1) 3D classification in C1 to remove the “missing” subunit particles (or maybe it is just less visible). 2) group the other classes and refine in D3 with symmetry relaxation to see if that subunit is still ‘different’ from the others.
Is there any other workflow that might be useful in this situation?

Side question: if I have a map refined with symmetry (say in D3) but with symmetry relaxation applied, which symmetry should I indicate in the deposition?
Thanks
Stefano

2

I would recommend:

  1. Refine in D3
  2. 3D classification (C1) using the D3-refinement output (with a mask around the entire hexamer), at least 6 classes to allow for each of the 6 possible positions of the missing subunit. For compositional heterogeneity I would recommend a filter resolution of 10-15 Å. Then use align 3D maps with update particle positions on to re-align the particle sets from the identical classes (with different orientations), then proceed to local refinement in C1.
  3. As an alternative to (2), you can try symmetry relaxed refinement, starting from the output of (1). Just make sure to add a few extra cycles for proper convergence of the relaxation procedure.

If the map is refined with symmetry relaxation, I would deposit as C1, unless there is an option for Frealign notation, in which case it would be lowercase d3 I believe.

Hope that helps!

Oli

1 Like
3

Thanks Oli for the suggestions!

4

HI!
I tried symmetry relaxation, but it didn’t help much… I was also thinking of trying symmetry expansion followed by local refinement, but I’m not sure how to do that…
My idea was:

  1. symmetry expand particles from my best consensus reconstruction in D3.
  2. perform local refinement in C1.
    For local refinement, would it be better to use a mask that covers the whole assembly or just a single monomer? In this case, I am interested in finding differences between protomers in the hexamer, but also in having the best model of the monomer.

I will also try 3D classification as suggested by Oli, but I wonder if it better to use the original particle stack or the expanded one.
thnaks
Stefano

5

Try both whole complex and monomer masks for symmetry expanded work - and I’d do 3D classification before local refinement. :slight_smile:

6

Hi Stefano

Like I said, my first preference would be 3D classification after D3 refinement, no symmetry expansion, mask over the entire oligomer.

You can also try sym expansion, but then I would use a focus mask just covering one monomer.

Cheers
Oli

7

Thanks you all for the suggetion! I’ll try all these
S