Gain reference using

1

Greeting there,

My data were collected and saved as mrc format without gain normalized.
When I have motion correction running of cryoSPRC, I realized there is obvious white line in horizontal and vertical direction of my sum image. So I come to my facility and was told I need flip X and Y of my gain reference, but I realized no improvement after trying. So I tried different combination of flip and rotation, and found flip X gave best result but with horizontal line there still.
So here I want to confirm that, for frames save in mrc format (not tiff), in which I should use my gain reference?

Thanks and cheers,
Yanhe

2

mrc or tiff shouldn’t matter, as long as it is still a movie file. (you can figure this out by e2iminfo.py)

If your facility failed to provide how gain-ref should be applied, I suggest trying all combinations by yourself, using 4-5 micrographs to speed this test up. Note other than X- or Y-flip, there is still “Rotate gain ref” to consider. Second note, “Rotate gain ref” in cryoSPARC means “Rotate gain ref counter-clockwise by 90 degrees this many times”, so if the camera is K3, only 0 or 2 is possible for this number.

3

If the gain reference isn’t working, I’d suggest making your own. Symlink ~1,000 micrographs to a directory somewhere and run the cisTEM program sum_all_tiff_files (or sum_all_mrc_files, if MRC format). Then use the resulting gain.mrc, so you won’t have to worry about rotation or flipping.

4

Dear all,

Thanks a lot for your favorable reply, I’m trying it right now.

Thanks and have a great day,
Yanhe

5

Greeting there,

I’m trying to use sum_all_mrc_files of cisTEM, after resume it, I only see options like Output sum file name, Estimate Dark and Gain images, Output dark and gain file name and threads to use, but could not find where to give directory.
Could you please kindly point me to right place?

Thanks and cheers,
Yanhe

6

I believe it just outputs the sum in the current directory. You can also use relion_estimate_gain for the same purpose. I would try to first remove any images that have significant regions of carbon or gold (or other high contrast contaminants) although probably not strictly necessary if you only have a few.

Cheers
Oli

7

Oli is right; output location is the current directory.

relion_estimate_gain also works fine, but you need to give it a movie STAR file input, which means a little more prep work.

1 Like
8

Greeting there,

Thanks a lot for favorable reply. I will have cisTEM try firsty. I’m trying to run sum_all_mrc_files, but have no idea how to give the directory for the program in which movies located

Thanks and cheers,
Yanhe

9

Just run it in the directory where the movies are located (or symlink the movies to the directory where you want to run the averaging)

10

Greeting there Oli,

Thanks for favorable response. Yes, it works.

Thanks and have a great day,
Yanhe

1 Like