IsoNet for non-cryo EM tomography

[Fiona Sievers]

Hi IsoNet Community, 

We are working with a TEM tomography dataset of approx. 200nm-thick resin-embedded samples generated using dual-tilt on a 200kV FEI TecnaiG2 TF20 (field emission gun microscope). We reconstruct these two-tilt tomograms using Etomo in IMOD (which contains a SIRT step) and wanted to use IsoNet for missing wedge correction. We have run into three issues and were hoping for your insights. 

We are using the deconvolve command to deconvolve the reconstructed tomograms. Since our voltage is at 200kV instead of 300kV, we tried changing the "voltage" parameter in isonet.py deconv (call: isonet.py deconv Tomo.star --snrfalloff 0.2 --deconv_folder /deconv_test --voltage 200.0 --tomo_idx 2). However, we get the following error, which I believe is similar to Stefan's deconvolve error (posted 16th Nov, 2023): 

The following modules were not unloaded: 

 (Use "module --force purge" to unload all): 

1) StdEnv 

 'voltage' not recognized!

Additionally, we changed the snrfalloff value to 0.2 for deconvolution and the training iterations to 20 based on trial and error for which values seemed to work best for our dataset visually and were wondering whether there is a better approach to finding the ideal snrfalloff, deconvstrength, and iteration values than visual inspection? Additionally, do you have any other general advice to applying IsoNet to transmission electron microscopy? For reconstructing our tomograms, would a weighted back-projection instead of SIRT be more compatible with IsoNet?

Lastly, we have some samples with gold beads embedded inside of them (not on the surface, on which we "delete" the gold beads prior to IsoNet using Etomo). When we run these tomograms through IsoNet, the circular shape of the gold beads is deformed, adding a halo around the circular gold. What could be the reason for this and is there a way we could get rid of this artifact? We have attached a picture with different brightness/contrast settings of the non-circular gold below. 

Thank you, 

Fiona 

[YUNTAO LIU]

Hi Fiona,

I am currently out-of-town and will get back to you later about the error and artifacts if no other people have any better suggestions.

I do not have too much experience on the resin embedded sample. I believe the parameters of deconvolution are only empirical, using wrong voltage should now affect anything. You can test until it gives you best contrast and details through visual inspection. IsoNet refine results in general have similar contrast of your original tomogram. There is no preference of WBP over SIRT. 

--
You received this message because you are subscribed to the Google Groups "IsoNet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to isonet+unsubscribe@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/isonet/d2f12ff2-0d4c-4705-86db-eeb16287a0c6n%40googlegroups.com.

--

Best Regards,

Yuntao Liu,  Postdoc.

California NanoSystem Institute

University of California Los Angeles

[Fiona Sievers]

Hi Yuntao, 

Thank you for your answer, that was very helpful!

Looking forward to your (or the community's) input on the other questions. 

Thanks and best, 

Fiona 

To unsubscribe from this group and stop receiving emails from it, send an email to isonet+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/isonet/d2f12ff2-0d4c-4705-86db-eeb16287a0c6n%40googlegroups.com.