Dear CryoSPARC user community, I would like to confirm that an approach we take is in fact the correct and best one. We work on a weak dimeric complex (AB). We have A, B, and AB in our sample. From 100.000 particles (particle-set1) we get a reconstruction centered on A. However, B moves relative to A and we only see a featureless cloud for B. We subtracted A from the micrographs, centered on B, 2D classified B to remove particle images of A alone, and have now a decent reconstruction of B alone (particle-set2, about 10.000 particles).
Now that we know which subset of particles contain AB (particle-set2) we want to go back to the original particle images (without subtracted A and centered on A) and reconstruct A from the subset that contains also B. In other words, we want to identify the subset of particles in particle-set1 that were used to reconstruct B (the ones that match particle-set2).
The way we did this, was to run a homogeneous refinement and use particle-set2 as input (the subset containing AB) but instead of particle-set2.particles.blob we use particle-set1.particles.blob. The resulting reconstruction seems to come from the expected subset, and is centered on A. Is that the best way to do this?