handedness (tilt angle and defocus)

[Jesse Hansen]

hi all,

I am thoroughly confused about handedness in WARP. 

I have two different pipelines:

1) WARP > imod > WARP > RELION

2) WARP > AREtomo > WARP > RELION

for pipeline #1, I have not been clicking "Don’t invert tilt angles" when reimporting after imod.  Is that the correct decision?

for pipeline #2, should I click "Don’t invert tilt angles" when reimporting? 

Secondly, if the defocus handedness ("check tilt handedness") of a tilt is indeed inverted (a negative correlation) then I will flip the handedness.  When I do this, does this automatically flip handedness for every tilt series I have loaded?  If so, I should re-process ALL tilt series CTF estimation, not just the current one, right? 

Is it possible that only some tilt series have incorrect defocus handedness? I recall in previous datasets that most tilt series had correct handedness but about 10% did not.  I went through and flipped handedness for those, one at a time. I am now realizing that probably makes no sense since flipping for a single tilt series will flip ALL of them.

Lastly, what is the relationship between the tilt angle handedness and the defocus handedness?  If I modify one then should I not modify the other?

thanks!

Jesse

[Alister Burt]

Hi Jesse,

This is a confusing topic!

There are two separate concepts here

1. physical handedness

2. defocus handedness

Physical handedness is what you’re probably used to thinking about, does the tomogram/the particles exactly match the reality of the sample we put in the microscope or is it a mirror image?

Defocus handedness is a little tricky. In Warp we end up with two defocus estimates

- a per movie, spatially resolved (2x2) estimate from frame series processing

- a per tilt, not spatially resolved estimate from tilt series processing

When reconstructing a particle at a specific position in the tomogram we take the per-tilt defocus estimate and apply an offset based on where that particle is along the projection direction. If the particle is above the center, we shift the defocus a bit one way, if it’s below we shift the defocus the other way.

Exactly which direction we apply this offset in is the defocus handedness. You can determine whether the defocus handedness is correct by looking at an image with a known tilt angle and in plane rotation and checking that the defocus changes across that image in the way you expect, this is what the defocus handedness check is doing.

As per the docs, warp has a different internal convention about what tilt angles mean, so angles should end up inverted relative to imod internally. I agree the way this is presented in the GUI can be confusing. I believe aretomo follows imod conventions.

You will get a better tilt series CTF estimate if the defocus handedness is correct as the model it fits takes into account how the ctf should change across images in the tilt series.

You could end up in a situation where some tilt series have correct defocus handedness and others are incorrect if your in plane rotation angles are inconsistent when importing tilt series. It’s more likely not the case though.

Defocus handedness only depends on physical handedness in warp because physical handedness is often changed by inverting tilt angles. This results in a flip in the XY plane. If you had flipped along YZ instead you would have a physical handedness change but no defocus handedness change

Hope this helps,

Alister

Sent from mobile - apologies for brevity

On Feb 16, 2024, at 02:38, Jesse Hansen <bluu...@gmail.com> wrote:

hi all,

--
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 on the web visit https://groups.google.com/d/msgid/warp-em/69cacf0d-32a3-4b44-9474-1659e65e6554n%40googlegroups.com.

[Jesse Hansen]

hi Alister!  

thanks a lot, I appreciate the explanation.  I think I now understand conceptually the difference.

However, from a practical point of view I'd like to ensure I'm processing correctly.   When importing to tilt series I will not click "Don’t invert tilt angles" since AREtomo and imod use the same convention.  Also, in what conceivable scenario would I want to click that box? 

As a second point, my workflow has been: immediately process all tilt series upon reimporting after WARP then check each TS individually for ctf handedness and flip as needed. In hindsight that makes no sense.  Instead I should import, then go through a bunch "process only this item" and "check ctf handedness" to get an idea of whether the entire collection of tilt series have inverted handedness.  If it's inverted, then do "flip handedness", which will flip it for ALL in the collection.   Only at that point do I process the CTF for the entire collection in the overview tab.  Make sense?

thanks again,

Jesse Hansen  

[Alister Burt]

Some people like to ensure physical handedness is correct in their tomograms, this button gives them a way to flip physical handedness

Makes sense!

To view this discussion on the web visit https://groups.google.com/d/msgid/warp-em/4fd4e6ec-00b8-4bd8-8fc1-0c8abb5254f7n%40googlegroups.com.