Hello, I am new in protein reconstruction and like Cryosparc though have a lot of questions!
Could you please recommend what steps to use if I have several similar regions in my helical filaments and only 1 region that should be different from the others? When I do helical refinement, Cryosparc averages all regions and I don’t see local changes.
Hi,
It’s hard for me to tell what the problem is based on this post. Could you elaborate a little more. Here are some thoughts that may or may not be helpful.
By definition helical symmetry means that there are multiple copies of the same asymmetric unit. Are you saying that there are multiple proteins/domains in your helical filament are being averaged together or that you are unable to resolve local variations in your protein’s conformation using helical reconstruction?
Some thoughts:
- Is the helical symmetry being imposed correct?
- Are there local variations in your filament subunits that you are trying to resolve?
-If you have correct symmetry but point 2 is your problem try something like relaxing the symmetry using homogenous refinement from your starting density map or 3DVA analysis.
I would be happy to share more thoughts/suggestions I just need to know more.
Best,
Mark Kreutzberger
Postdoc, Egelman Laboratory
Hello, Mark!
Thank you so much for your reply! Nice to meet you! I am a Postdoc at BU.
I use C1 symmetry, I know the twist and rise, and the structure looks like pretty well, resolution 2.7 A. But. The task is to find local variations in my filament. Roughly, I have 7 pseudo-similar domains along the filament. I suppose at least 2 of them should be kind of different, but remain too similar for Helical refinement.
So, do you suggest to use Helical Ref —> then Homogenious Ref? I tried to do 3DVA but 600 box is too big and it did not work after 16 h of processing, I am re-extracting to 400 box. In ideal, to classify the domains to 7 classes… but I can’t figure out the workflow.
Also, I don’t understand why Ab-Initio Rec does not work after Filament Tracer (it averages me oppositely directed repeated parts of the filament), and Heterogen Ref works after Ab-Initio Rec only and not after Helical Ref.
With best wishes, Olga
Hi Olga,
Nice to meet you as well! Are the 7 pseudo-domains repeating? If so, it should be pretty easy to resolve them by relaxing the symmetry. If, not 3DVA might help. If they are repeating, is the helical symmetry you are using assuming that the 7 domains are in a single asymmetric unit (ASU) or are the 7 domains spread across several ASUs. It might be possible to define a symmetry to include the 7 domains into a single ASU, if they are not already.
The simplest thing would be to try to use helical refinement and then do homogenous refinement, filtering it to a low enough resolution that the homogenous refinement converges on a reasonable structure but it doesn’t skew your results. 20-30 Å might work.
Ab initio doesn’t work on every type of assembly. For bacterial flagellar filaments I have never gotten it to work. Some things to try playing around with might be the starting ab initio job resolution. Heterogenous refinement doesn’t always work for me either, I’ve yet to really play around with it significantly.
My success in 3DVA jobs really depends on the mask (softer is typically better) as well as the resolution I am filtering to. If your differences can be spotted at low resolution, you can try binning your particles. You can also try reconstructing your particles with a smaller dataset and then running 3DVA.
I hope at least some of this was helpful!
Best,
Mark