Hi
I am doing reference-based motion correction and I am getting CV score concentrated at 0 trajectory. I used Minimize over per-particle scale during homogenous refinement as suggested in tutorial video. earlier when I was not minimizing per-particle scale, this graph was looking as in the example of RBMC. is this a problem and how can I fix it?
Thanks
Dhiraj
@dhirajks What are the format and dimensions (pixel and frame counts) of the input movies? Please can you also post a GSFC curve showing the current state of volume refinement?
the input movies are in tif format (K3), 40 frames. the GSFC curve is attached. I used local CTF, then Global CTF before doing RBMC. Also I fourier crop to box size of 200 pixel and I extracted particle in 400-pixel box. data were collected in superresolution mode, and my motion correction was already done with 2X binning.
Hi @dhirajks,
It appears that a set of particles have a trajectory > 1e13Å with that point in the upper right. Could you use a Curate Exposures job and sort by the Motion Distance column to see whether there are any outliers (example here of what you might find)?
Additionally, can you provide a plot of the PPS in the upstream homogeneous refinement job?
Best,
Kye
