Recenter and rebox after subtraction

Local Refinement
1

I am wondering if it is possible to add the functions of “align 3D volume” and rebox into particle subtraction. For particle subtraction, we sometimes only want a small part of map and would like it to be in the center and with smaller box.
Thanks,
Lei

2

Hi @rainfieldcn,

Thanks for the query – the recommended way to accomplish this in CryoSPARC is by performing the steps separately. For example, one possible workflow looks like:

EDIT: added two tweaks from @Hui (thanks!)

  1. Volume alignment tools (calculate mask center of mass):

    • Supply a volume, particles, and a mask covering the part of the map of interest; re-center to the center of this mask
    • This job will report the center of mass of the mask and can be used in the next job.

  2. Volume alignment tools (re-center):

    • Supply the same volume and particles as in job 1. Do not input any mask. Use the calculated center of mask as the 3D coordinates of new center.

  3. Extract from micrographs (re-extract)

    • Re-extract particles from Step 1 or 2 with “re-centering based on aligned shifts” enabled. Pick a box size large enough such that the particles are not cut off or too close to the edge. You can adjust the box size if it is too small after inspecting the output volume from step 4.

  4. Homogeneous reconstruction only

    • Use particles from step 3. Ensure that the output volume is not too close to the box edge

  5. Create a mask around the area to subtract from the output volume from Step 4

  6. Particle subtraction (subtract)

    • Connect the inputs below and run the job:
      • mask from Step 5,
      • particles from Step 3, and
      • volume from Step 4

  7. Downsample particles (crop)

    • Set the real-space crop size to a reasonable number, based on the desired final refinement.

Best,
Michael

2 Likes
3

Hi Michael,
Thank you for reply!
Is Step #5 carried out in Step #2? Or is there another function for this cropping process?
Best,
Lei

4

Dear @rainfieldcn,

In step 2 you have the choice to choose a new box size. This box size must be big enough to cover both the centered region of interest, as well as the region that is being subtracted in step 4. In fact it should still be at least 2x the size of the full protein (as per our standard recommendation of the box size). This is because particle subtraction uses the CTF of the subtracted region, and if this region is near the edge of the box, signal can be cut off (see this reference for some discussion). It is only after particle subtraction occurs that we would recommend cropping to the final smallest box size, via downsample particles

Best,
Michael

5

Hi @mmclean,

Thank you for summarizing the necessary steps here, but I noticed a couple of potential issues of the workflow that i want to share here.

  1. The volume output of step1 is masked. So to get an unmasked, but translated volume for subtraction, we have to repeat step1, but this time without a mask, using the calculated number of mask center. Can cryosparc also output the unmasked volume here so we don’t have to repeat?
  2. If the particle box is not large enough, then the shifted volume from step 1 will be truncated. In this case, we have to do a reconstruction job using particles from step2. Therefore the box size of the particles in this step has to be big enough.

Best,
Hui

1 Like
6

Dear @Hui,

Thanks for the report. You are right on both accounts, my apologies for the error!

Good request, you’re right that currently you have to create a new job without the mask supplied and with the manually entered re-centering coordinates. We have this recorded now to allow volume alignment tools to allow outputting the unmasked volume.

This is correct. I have updated my comment to add these two notes.

Best,
Michael