Problem with import on a cryoET dataset

[Andrea Dallapé]

Hello,

I'm working on a cryoET dataset. The frames have been motion corrected in Warp and resulted in tiff files. I'm trying to move on Eman2. I started by trying to build the stack:

e2buildstacks.py micrographs --tilts --guess

(with the tiff files in the micrographs folder), but I quickly realize that it does not work with tiff files. So I used 

e2proc2d.py *.tif @.mrc

to convert the tiff in mrc. The e2buildstacks command works, creating a .lst file in the tiltseries folder, as expected:

input is a directory. reading all mrc/mrcs/hdf files in it...
found 41 files
File name of the first input:
micrographs/JTL018A_2_lam1_ts_006_001_0001_-0.0_22.39.24.mrc
11 Columns
18,  2,  1,  6,  1,  1,  0,  0,  22,  39,  24
Guess column 6 is for tilt angles,
ranging from -51.0 to 69.0.
JTL018 : 41 images -> tiltseries/JTL018.lst

I'm trying at the moment with a single tilt series. When I try to import the tilt series to create the hdf file:

e2import.py tiltseries/JTL018.lst --apix=1.66863 --import_tiltseries --compressbits=8 --importation=copy

this error occurrs:

e2proc2d.py tiltseries/JTL018.lst ./tiltseries/JTL018.hdf --inplace  --apix 1.66863  --compressbits 8
Input file 'tiltseries/JTL018.lst' does not exist.
Done.

that results in the deletion of the .lst file in the tiltseries folder. This happens both through the command line and the GUI.

Can you help me by pointing out what is wrong and how to correct it?

I'm working on a Mac M4.

e2version.py
EMAN 2.99.66 ( GITHUB: 2025-04-11 16:08 - commit: cfd9aefa6 )
Your EMAN2 is running on: Mac OS 15.3.1 arm64
Your Python version is: 3.12.10

Thank you for any help and kind attention.

Best,

Andrea

[Muyuan Chen]

Normally if you built the tilt series through e2buildstack, you shouldn't need to import. If you just rerun the buildstack command  (double check that you can open the lst file in e2display browser and pixel size displays correctly), and skip the import step, can you reconstruct the tomogram properly using the lst as input?

--
--
----------------------------------------------------------------------------------------------
You received this message because you are subscribed to the Google
Groups "EMAN2" group.
To post to this group, send email to em...@googlegroups.com
To unsubscribe from this group, send email to eman2+un...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/eman2

---
You received this message because you are subscribed to the Google Groups "EMAN2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eman2+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/eman2/CALnnSmfxtES54Lj37_iyfAsuN_J0Z71aeamduFfWt5eyrYgCZQ%40mail.gmail.com.

[Andrea Dallapé]

Hi Muyuan,

From e2display (Show 2D) the stack seems good.

I thought the same, so I continued with the tomogram reconstruction

e2tomogram.py tiltseries/JTL018.lst --zeroid=-1 --tltstep=3.0 --npk=20 --tltkeep=0.9 --outsize=1k --niter=2,1,1,1 --bytile --pkkeep=0.9 --compressbits=8 --clipz=320 --bxsz=32 --filterres=60.0 --rmbeadthr=-1.0 --threads=12 --patchtrack=2 --autoclipxy --extrapad

(tried with tltkeep=1 and/or --clipz=-1). The

But the result is extremely bad. This tilt series was previously reconstructed in IMOD with very good results. I was thinking it is because of the tiff files/conversion...

Do you have any suggestions?

Best,

Andrea

To view this discussion visit https://groups.google.com/d/msgid/eman2/CAO_0xqFTcpf9_O%2B%2BpKW0TffCcAV1eaR345UVZzYODwUGth3sYA%40mail.gmail.com.

[Muyuan Chen]

It is hard to guess without seeing the images. One possibility is buildstack sort the images incorrectly. When you display the 2D images do you see a continuous tilt series? 

Also check the temporary files to see at which step it fails. Is the tiltseries_transali smoothly aligned? Is the tilt axis estimation correct? Or does it go wrong after the landmark based tracking?

Muyuan

On Apr 16, 2025, at 12:59 PM, Andrea Dallapé <dallape...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/eman2/CALnnSmet8w3V380jNJBhOuni-5K8OXrmyCnPgpH%3DDQKb%2Bd5mLw%40mail.gmail.com.

[Andrea Dallapé]

Yes, the tilt series is continuous, I don't have any doubts about buildstack.

The reconstruction is correct, is the quality of the visualisation that is concerning me. Could be uniquely because it is binned 4?

I tried to reconstruct again using the tilt axis suggested by the CTF handedness:

e2spt_tomoctf.py tiltseries/JTL018.lst --dfrange=2.0,7.0,0.02 --psrange=10,15,5 --tilesize=400 --voltage=300 --cs=2.7 --nref=15 --stepx=20 --stepy=40 --checkhand --threads=1 --writetmp

Average score: Current hand - 0.399, flipped hand - 0.390
Defocus std: Current hand - 1.060, flipped hand - 0.986
Current hand is better than the flipped hand in 48.8% tilt images
tiltseries/JTL018.lst: Result: the handedness seems to be flipped. Consider rerun the tomogram reconstruction with --tltax=-270.9 then rerun the CTF estimation.

But the hand better is far below 80%...

Here the comparison between imod and eman (using the command as before, with --extrapad and --niter=2,1,1,1). I wanted to use the convolutional neural network particle picker to pick ribosomes, and I cannot use the reconstruction from imod (or something changed later?). Moreover, the reconstruction in eman is suffering from some kind of visual artifact, similar to patches...

Do you have any idea?

Best,

Andrea

To view this discussion visit https://groups.google.com/d/msgid/eman2/98928C78-3672-4C22-A466-C4AAA3032AFD%40gmail.com.

[Steve Ludtke]

The tiling artifacts can be corrected with either or both of the --moretile or --extrapad options. They tend to be necessary with thicker tomograms, and in some cases also indicate that an insufficient Z thickness is being reconstructed (--clipz). If you look at your reconstructed tomogram in profile (X-Z or Y-Z) you should be able to see the full thickness of the specimen. Particularly in higher magnification cellular tomograms the default is insufficient.  Use --help to see the full list of options

To view this discussion visit https://groups.google.com/d/msgid/eman2/CALnnSmcddz3YEWe9RTWZ4HC7_YE-X_cu6gGuWBmp17pKT6yZuQ%40mail.gmail.com.
<eman_zoom.png><imod.png><eman.png>

[Andrea Dallapé]

Hi Steve,

sorry for the late reply and thank you for the suggestion.

I tried to use --moretile in addition to --extrapad, and the tiling artifacts disappeared. Regarding --clipz parameter, can you kindly explain better how I should change it? If I have a 220 nm lamellae, make it sense to change from -1? In this case, what value would you suggest?

Thank you so much for your help.

Best,

Andrea

To view this discussion visit https://groups.google.com/d/msgid/eman2/3C970EE0-5108-4EC9-AA66-960EED78AA48%40gmail.com.

[Steve Ludtke]

if you have a 200 nm lamella, you would normally want to reconstruct 1.5 - 2x thicker than this along Z, so say 300 nm. If you are generating default 1k x 1k (4x downsampled) tomograms and your full scaling A/pix is (for example) 2 A/pix, then at 1k x 1k it will be 8 A/pix,

300 nm/0.8 nm/pix = 375 pixel targeted Z thickness. You should always use a “good” size for any dimensions, so round that up to 384  (e2help.py boxsize, which would be  --clipz=384

This is only a little larger than the default clipz of 5/16 of the size (1024 x 5/16 = 320 pixel default). So, for a 200 nm lamella at 2 A/pix you might be fine with just the default, but if you are at higher mag or have a thicker lamella, then you might want to adjust it. 

To view this discussion visit https://groups.google.com/d/msgid/eman2/CALnnSmdBRby1SEDX6NWF7O_Lb-wa8Po5t219JAeqPJiMnBZi0w%40mail.gmail.com.