refine_stageangles

[P C]

Hi all, 

What does --refine_stageangles switch do exactly? In particular, does it refine individual tilt angles (per tilt series)? If it does, the logic tells us it should refine tilt shifts as well. 

If this is the case, this option unlocks Emclarity/Relion-style frame alignment.

Somewhat related to this: what would be the meaning of --refine_imagewarp 1x1? Does it amount to full frame XY shift refinement? (In which case combining it with stage angle refinement could help to fine-tune tilt series alignment). 

Ultimately, I am looking for the best ways to improve the entire tomogram reconstruction, which could allow us to identify finer details of a biological assembly. What would be the best way of doing something like that in Warp/M?

Peter

[Alister Burt]

Hi Peter,

Your intuition is correct, M optimizes its model of the entire multi-particle system which includes (for tilt series)...

- particle poses (optionally resolved over time)

- deformations (image/volume warp)

- electron optical params (defocus + higher order ctf stuff)

- stage tilt angles

- probably more things I'm forgetting right now

Rigid body 2D shifts of tilt images are captured by a 1x1 image space warping as you suggest - I recommend reading the Warp and M papers carefully, especially the methods sections, they're a goldmine and very clear.

You can find some discussion started by Euan Pyle who works at EMBL on the topic of improving tomograms in this thread

https://groups.google.com/g/warp-em/c/0rdmC9CU1L4/m/bQ5LzLyhAAAJ

Cheers,

Alister

--
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/5985a0c0-50d2-4f39-b597-6e0d5e1ccbdcn%40googlegroups.com.

[P C]

thank you!