Can we use reconstruct3d to reconstruct cisTEM par files? I've tried using the imported stack and an output .par, or exporting to Frealign, but I just get a NaN volume. I've also tried normalizing particles or not, inverting contrast or not, etc., and the output indicates normal termination. From reading the code so far I haven't figured this out.
I mainly want to reconstruct maps using your reconstruct3d, but not normalizing so I have the direct particle intensity scale.
Yes, this should work as long as the particle stack and par file match up. Are you saving the par file from the refinement results panel? Also, are you inputting the radius for mask, not the diameter?
Thanks for the reply, Tim. It turns out I *had* been giving the diameter even though my own variable was named radius. It's behaving as expected now.
You mentioned subtraction will be available in a future version of cisTEM - can I ask how you are scaling the references / projections for subtraction? Even with normalization set to "no," the intensity scale is very different compared with relion_reconstruct. When I make projections from the relion output, the intensity scale matches the particles (on average).
Yes there is some additional scaling required to get things to match up. I can't remember off the top of my head what it is, it is probably something like sqrt(number_of_pixels). If you need it, I can tell you in more detail?
If it's not too much trouble, I'd appreciate it!