ctf correction in half tomograms?

42 views
Skip to first unread message

giulia.z...@gmail.com

unread,
Apr 29, 2026, 8:30:46 AMApr 29
to Warp
A quick question. While I understand the warp reconstructions are 3D CTF corrected by default, I wanted to to double check that half tomograms (generated with --out_average_halves during fs_motion_and_ctf, followed by --halfmap_frames in ts_reconstruct) are also CTF corrected?

This would important to check for running IsoNet2.

Thanks

Giulia

Huy Bui

unread,
Apr 29, 2026, 10:15:44 AMApr 29
to giulia.z...@gmail.com, Warp
Perhaps, this can be answered better by Dimitrov, but I believe the half maps are CTF corrected.

My anecdotal evidence is based on the fact that I did my test by comparing DeepDeWedge with Warp half maps vs IsoNet2 (Correct CTF or ignore CTF). I saw that DeepDeWedge results look as good as IsoNet2. It indicates that the advantages of CTF correction in IsoNet2 vs. DeepDeWedge diminished if we fed DeepDeWedge a CTF-corrected tomogram.

Best,



--
You received this message because you are subscribed to the Google Groups "Warp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to warp-em+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/warp-em/d83e7d44-2732-48e7-a047-9f36c5ce832bn%40googlegroups.com.

Giulia Z

unread,
Apr 29, 2026, 10:23:59 AMApr 29
to Huy Bui, Warp
Thanks - I asked Claude to look into the source code and it confirmed that the half maps are indeed ctf corrected, but I was hoping for human confirmation…

Giulia

teg...@gmail.com

unread,
Apr 29, 2026, 10:46:45 AMApr 29
to Warp
Human comrade Dimitrov reporting for duty ;-) Yes, they're all produced by exactly the same code, just feeding it tilt movie averages or half-averages. I would fully trust robot comrade Claudov with code-defined answers like this.

Karine Mazmanian

unread,
Apr 29, 2026, 10:51:52 AMApr 29
to teg...@gmail.com, Warp
:-)) i wanted to jump on this train and ask if it is possible to omit the 3d ctf to generate tomograms? I wanted to try out cryoDRGN for some of my datasets, which i already processed with WarpTools. Any insights are appreciated.

Karine

teg...@gmail.com

unread,
Apr 29, 2026, 12:25:54 PMApr 29
to Warp
Hi Karine,

You could modify the XML metadata to set the CTF parameters to amplitude = 1, Cs = 0, defocus = 0, astigmatism = 0 (the last 2 have values for each tilt) to get a flat 1 in the CTF to effectively disable it. If you don't want to touch the metadata directly, you can use https://github.com/warpem/warpylib to open, modify, and save the metadata in Python. You can also use warpylib directly to obtain the reconstructions and not touch Warp at all. Warpylib would also be easier to modify if you want to do something completely custom.

Cheers,
Dimitry

Karine Mazmanian

unread,
Apr 30, 2026, 2:48:59 AMApr 30
to teg...@gmail.com, Warp
Hi Dimitry,

That’s a great solution! Thank you so much for your help.

Best,
Karine


Reply all
Reply to author
Forward
0 new messages