Dear community,
I was rescaling my dataset after extraction to fit the whole stack into RAM.
After 2D classification , initial refinement and sorting I want to use the unbinned particles, which I also extracted.
Is there a feature implemented to do so?
Best,
T.
Hi @tarek,
Presumably you used the “Downsample particles” job to Fourier-crop your particles before doing initial 2D classification/refinement etc, and now you’ve identified a subset of particles.
To now complete a refinement using the un-binned (pre-downsampling) particles but only using the subset that you’ve identified, the simplest option is to just re-extract the subset of particles (using the particle extraction or local motion correction jobs) and refine those.
Technically cryoSPARC v2 does retain the correspondence information between the binned and unbinned particles, and in an upcoming version we will be deploying a UI upgrade that will allow you to connect the subset of binned particles + the unbinned particle stack to any other job and cryoSPARC will automatically take the intersection of those sets, yielding the subset of particles but unbinned images.
Hope this helps,
Ali
Hi Ali,
thanks for the helpful reply.
An UI upgrade in the next version would be great, however your suggestion does the job well.
I usually start with the extraction of unbinned particles and create binned stacks based on them (makes old-fashioned bookkeeping easier).
Following the currently implemented workflow unbinned particles should preferrably be extracted at a later step, right?
Best,
T.
That’s correct, though depending on how you’re doing the extraction (did you import micrographs after doing motion correction outside cryoSPARC, or are you doing local motion correction within cryoSPARC?) you will need to extract the unbinned particles, and then downsample them, process those, then extract again (unbinned) only the subset you like.
Ali
Hi Ali,
I have a related question on that topic. Is it correct that when importing K2 super-resolution movies into cryoSPARCv2 to do global and local motion correction within cryoSPARC I can only bin my data (using the Downsample particles job) once I extracted them with the super-resolution pixelsize using “Local motion correction”? This leads to quite big initial particle stacks. Are you planning to add an option to bin the data after global motion correction, before local motion correction/extraction? Thanks.
Simon