About an error in "Find particles from segmentation"

[毛翘顺（emmao）]

Hi!

I am using the segmentation function in EMAN2 and it seems like the segmentation works well for nucleosomes in my data as I opened the .hdf file in chimeraX after "apply the neural network" section. Below is the result:

You can see that the density is clustered at two sides, the boundary looks like a "W". However,  when I tried "find particles from segmentation" to export the 3D coordinates of those particles, (I just import the .hdf file of the imported tomogram and the segmentation result which i opened in chimeraX, other parameters were left default ) the log says 18799 boxes found

but when I opened the .hdf file of the imported tomogram in "manual boxing" it seemed that the boxes were located at the "opposite" region, nearly none of the particles in the "W" boundary were boxed. The particles boxed here were totally different from the .hdf file after "apply the neural network" and opened in chimeraX.

Thank you for any suggestion.

[Steve Ludtke]

The "particle extraction" I believe works on connected densities from the segmentation. Your segmentation looks like it's pretty noisy for a method like that. If you want to go that route, you might want to low-pass filter your tomogram first?  

Another alternative would be to use the "convnet based autoboxing" tool which is targeted at particle picking rather than annotation, and does a very nice job for most particles.

We also have "e2tomo_annotate.py" which is still under development, but is usable, and provides a complete GUI for multi-feature annotation/picking. 

It looks like the version of EMAN2 you're using is _very_ out of date, since "convnet based autoboxing" doesn't even appear in the subtomogram averaging submenu in the GUI snapshot you sent. As long as you have a conda-based install, you can install a newer version extremely easily following this:

https://eman2.org/Install/CondaInstall

and if you don't want to mess with the existing install, you can just change the name of the environment it uses, and have both versions installed at once, eg:

conda create -n eman296 eman-dev -c cryoem -c conda-forge

On May 17, 2025, at 12:14 PM, 毛翘顺（emmao） <zcb...@ucl.ac.uk> wrote:

Hi!

I am using the segmentation function in EMAN2 and it seems like the segmentation works well for nucleosomes in my data as I opened the .hdf file in chimeraX after "apply the neural network" section. Below is the result:

<6feb7518915ab08f97a5385b66d40cc.png>
You can see that the density is clustered at two sides, the boundary looks like a "W". However,  when I tried "find particles from segmentation" to export the 3D coordinates of those particles, (I just import the .hdf file of the imported tomogram and the segmentation result which i opened in chimeraX, other parameters were left default ) the log says 18799 boxes found<15a107b8a1275285d7907018fc5749a.png>

but when I opened the .hdf file of the imported tomogram in "manual boxing" it seemed that the boxes were located at the "opposite" region, nearly none of the particles in the "W" boundary were boxed. The particles boxed here were totally different from the .hdf file after "apply the neural network" and opened in chimeraX.

<e062cd5bb12686b174db21c66a6648e.png>

Thank you for any suggestion.

--
--
----------------------------------------------------------------------------------------------
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/eman2/37ee4597-4d8a-453a-be06-043d35e969f1n%40googlegroups.com.
<15a107b8a1275285d7907018fc5749a.png><6feb7518915ab08f97a5385b66d40cc.png><e062cd5bb12686b174db21c66a6648e.png>

[毛翘顺（emmao）]

Hi Steve, thanks a lot for your reply. I've already got the new version of EMAN installed and I will try the convnet based autoboxing! Thank you again for your suggestion!!