Leginon

We have observed the same for a while but could not find a software/hardware reason for it. The same is observed when moving with iterative stage movement to the multiple targets selected on an hl image. Even after the drift monitor result of the focus target declares an insignificant drift, the first exposure image would move more at first (based on drift tracking in movie-capable camera, and found in the first two frames, but more pronouced in the first 200 ms frame) than subsequent frames and more than the first two frames of the later exposure within the same parent image.

Do you have data that says this was not the case previously ?

I have conducted some more tests to (A) see if the drift is reproducible and (B) to find the source:

I switched to using a test grid of negative stain TMV on continuous carbon-coated quantifoil. Dose was around 20e/A^2 and exposure time around 700 ms. Drift monitoring was set to a limit of 4A/s.

The drift is present in the first exposure (2 defocii series but both set to 3.5um) in several exposure positions:

I then tested 5 position exposures with 2 defocii (both set at 3.5 um):

Again, you can see the drift in some of the first exposures but not the second.

I also tried a series of exposure targets where exposures of carbon edges alternated with exposures of pure carbon regions. The drift was only present in the edge images.
Exposures in the centre of the hole also gave no discernible drift.

My conclusions so far are that:
*The drift only occurs when an edge is present.
*The drift only occurs in the first exposure.
*The magnitude of drift seems correlated with the amount of carbon in the frame.
*The direction of drift seems to be perpendicular to the edge.

I will be modifying my cryo collection strategy to avoid carbon edges in future but it's going to make CTF determination and real-time data quality assessments more difficult.

This appears to get into the area of beam induced movement and/or charging. Your observation is interesting because I often hear people say that a bit of carbon in the beam (not necessary in the camera image) is one way to minimize charging. As long as you intend to take multiple exposures in a hole, it is most likely that some carbon would be in the beam.

On the other hand, there are reports that the best way to minimize beam-induced movement is to illuminate the hole in such a way carbon in all directions touches the beam. For example, our highest resolution data are now taken with smaller c-flat hole (1.2 um) and only one image is taken per hole to follow this guideline.

That makes sense, thanks Anchi.

If I am going to avoid carbon then I will need to determine CTFs from particles. Is there a way in Appion to calculate CTFs from stacks of boxed particles?

I know it used to be possible in AceMan but that required MATLAB.

If not, I will try write a python script to take exported EMAN box files and determine the CTFs in EMAN.

Thanks,
Jason

Jason,

I asked around and no, current Appion does not calculate CTFs from stacks of boxed particles.