[3dem] RELION - Template picking with imported 3D reference

[Agus Zavala]

Dear community,

I'm trying to import a volume from EMDB into RELION4.0.1 and use it for
template picking on my dataset, I've already changed several parameters,
but it keeps finding zero particles. I think I may be making a mistake in
the process.

Some details:

I used these parameters with a 3D model obtained from this dataset, on a
100 micrographs subset, and it finds ~500particles

+Symmetry = I
+3D angular sampling = 3.7degrees
+Lowpass = 15A
+Highpass = -1
+Pixel Size in ref = 2.28
+In plane sampling = 3degrees
+Reference have inverted contrast = Yes
+Are reference CTF corrected = Yes
+Ignore until first peak = No
+Picking Threshold = 0
+Minimum Distance = 400A
+Max stddev noise = -1
+Minimum avg noise = -999

Then I downloaded a .map from EMDB, saved it as .mrc in ChimeraX, and
imported it in Relion as 3D Reference (.mrc). Also, I tried changing the
format using other tools (map2mrc and e2proc3d.py).
Then I used it for Autopicking on the same subset with the same parameters
as before, except a few:

+Pixel Size in ref = 1.375 (taken from EMDB, and checked in ChimeraX)
+Reference have inverted contrast = Yes/No (tried both)
+Are reference CTF corrected = No (also tried Yes, just to check)
+Lowpass filter = 15A/60A (tried both)

but it always finds zero particles. The correlation between both maps in
ChimeraX is 0.843, they're quite similar.

Any suggestion why I can't make it work? Am I making an error in the
conversion of the map? or the importing to relion?

best,
Agustin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20250220/112b5c67/attachment.html>