I am trying to sort out polarity issues with a protein bound to my filament. I have a polar filament with a binder that interacts in such a way that it should always point towards the same end of the filament. However, I suspect the filament particles are being incorporated into the reconstruction in both directions equally, which is a problem. I have made a tight mask around the filament and did a helical refinement, with the hope that everything will align in the correct orientation. I then expand the helical symmetry and do a 3D classification without alignment (and without a mask) to see which direction the binder protein is pointing. Surprisingly, it is clearly pointing in both orientations equally. To resolve this I use the volume alignment tools to manually flip around the particles for those volumes which point in the other direction (for “3d rotation Euler angles” field I use a value of 3.14, such that the filament is flipped 180 degrees), combine the particles, and do a refinement of the filament + binder. Hoping this will improve the map.
I am just wondering about the step where I do symmetry expansion of the helical parameters, is that necessary? I am having a hard time conceptualizing what symmetry expansion will do in the case of helical symmetry since, unlike point group symmetry, the resulting particle stack has the same number of particles as the input stack.
This workflow sounds reasonable — when you use volume alignment tools to rotate some classes/particles to align their polarity to the other classes, do the final volumes all appear properly superimposed? If so, it would make sense to use local refinement as the final refinement in your step (since it needs the initial orientations from the rotated particles), as the other refinement jobs ignore the initial orientations.
To answer your question, symmetry expansion for helical particles should still multiply the number of particles by some factor, which is given by the “Helical symmetry order” parameter in the symmetry expansion job. If it’s set to 1, that means that the identity transform is used and the job doesn’t actually do any expansion, it just preserves the input particles’ alignments. (If it’s set to an integer greater than 1, expansion will be done). Note that if the upstream helical refinement indicated that a symmetry order of 1 was used, that means that symmetry wasn’t actually applied in Fourier space. The helical symmetry order of the helical refinement can be overridden by the Maximum symmetry order to apply during reconstruction parameter in helical refinement; see more info here.
Hi Michael, Each of my protein units is adhered to a tube based on specific helical parameters. I want to know if cryoSPARC can implement symmetric extension for each protein unit according to helical parameters, and later I would like to do local refinement on the protein units rather than the entire helical tube.
Thanks in advance!
Jinying Ma