Hi,
In v2.12, there is an option to set symmetry during variability analysis, however it is stated in the tooltip that this option doesn’t do anything, and suggests the use of symmetry expansion. This is a little confusing, why does this parameter exist if it doesn’t affect anything (i.e. if C1 is used anyway)?
And could you perhaps elaborate a bit more on how cryosparc deals with symmetry expanded particles during 3D-VA, and what the optimal strategy is for analysis of symmetric systems using this tool?
Cheers
Oli
Hi @olibclarke,
The symmetry option is just there to remind people that if they have symmetry, they should generally always perform symmetry expansion before 3D Var. If you put in a symmetry other than C1, the job will fail with an error message that says to use symmetry expansion.
3D Var is computing eigenvectors of the 3D covariance of the particle images, assuming fixed poses that come from the input. When there is “symmetry” in a particle, this only applies to the consensus structure - any variability mode can break the symmetry. However, the variability mode (assuming the underlying symmetry is a true symmetry) should naturally occur in all symmetric versions of itself.
In this sense there are two kinds of variability modes: modes that are changes/motion within just a single subunit, and modes where there are changes/motion across the entire particle in a coordinated fashion. The first kind are modes where each particle image contains information about multiple positions along the mode (since each subunit is in potentially a different position along that mode). The second kind are modes where each particle image contains information about only one position along the mode (since the entire particle is in only one position) but due to the symmetry, the image could be used as information for all symmetric copies of that same mode.
So for the first kind of mode, symmetry expansion is best. In this case it’s also fine to create a mask around a single subunit, but this is not necessary (since using the subunit mask will make it impossible to find motions across the entire particle). For the second kind of mode, using symmetry expansion will mean that there are many copies of the same mode that can be found (i.e. imaging a symmetric molecule bending along its entire length, in one direction. There are equivalent copies of that mode where the molecule bends in the symmetric versions of that one direction). Symmetry expansion makes sure that every particle counts for each one of these copies, rather than just the single one with which the particle is arbitrarily aligned in the input poses.
The above is still somewhat of a work in progress, and there are other cases (pseudosymmetry?) that should be considered. It would be great to hear your thoughts!
3 Likes
Thank you Ali that was very helpful!
Oli
Also @apunjani - one more query - I notice there is a new “Highpass resolution” paramater in 3D-VA. Just so I understand, is this actually high-pass filtering the particles? And is this intended to remove contributions from things only contributing to variability at low resolution, e.g. micelles? If so, what values have you found useful - I guess 15-20Å or so?
Cheers
Oli
@olibclarke the new highpass option is not highpass filtering the particles, but rather adding a “highpass prior” to the 3D Var components limiting the amount of power they can have at low frequencies. This essentially ignores variability that is larger than a certain scale, as you said. I haven’t really tried it for micelles (where I think the best strategy is still to mask out the micelle), but I was finding that with a lot of smaller/membrane proteins there is a huge amount of “structured noise” present in the images at low resolutions. This could be from pancaked particles floating around at the air water interfaces, empty micelles above/below the particles, etc. Those were causing most of the variability modes to be full of just large blobs appearing and dissapearing, rather than actually probing the motion or flex of the molecule. So I found that in that case, turning on the highpass resolution helped a lot. In those cases I set the highpass around 20A with order 8 (order is the steepness of the filter falloff - these are butterworth not rectangular filters).
However, I subsequently also implemented coloured noise in the 3D var algorithm, which does nearly the same thing but automatically. The colour of the noise is estimated first from all the particles, capturing those large low-res structured noise sources. That noise model is then used during 3D var so the contribution of low-res terms are automatically weighted down since there is large noise at those resolutions. This turns out to generally work very well across the small particles I’m testing, though I did find at least one case of a larger molecule where it was slightly better to use white noise.
2 Likes
Thanks @apunjani! Actually, even in the presence of a coloured noise model it seems like the highpass filter makes a big difference to the result, at least for one small, flexible sample that I just tried it on after reading this. That explanation was very helpful!
Cheers
OIi
I also notice that the “highpass resolution” parameter is also present in 3D var display - is it doing the same thing there? That is, if we set that parameter during the analysis step, should it also be set during display?
Oli
Along these lines I was curious if there was a way to consider variability in a given asymmetric unit in the context of the whole particle. For example, I have a C2 particle which I expanded and applied an asymmetric mask, and ran local refinement and finally 3DVA. The output suggests there is secondary structural element which varies in length. I would like to know if the long and short states can exist simultaneously in a single particle ( or conversely if the short/short or long/long states exist). Because 3DVA has to be run with C1 symmetry , I am assuming that each asymmetric unit is treated as its own particle and the relationships between subparticles are not considered. Is there an easy way to perform this analysis in cryosparc? I suppose I could model the three possible states and use them as seeds for a heterogeneous refinement, but it would be interesting if there was a way to consider higher symmetries in a 3DVA job.
1 Like