Are there limitations in sample preparation which might affect 3D reconstruction? For example, imaging a sample which has been stained with uranyl acetate and imaged without flash freezing methods.
Welcome to the forum. 
Yes. To use the case you mention as an explicit example, negative stain imposes hard limits on attainable resolution. It’s usually considered to have a maximum somewhere in the 15-20 Angstrom range, although this will vary depending on the stain used as the grains are different sizes. Also, how negative stain grids are imaged imposes a limit - the stain settles around a particle, creating a thicker “halo” region around the protein, with a thinner layer over the top, which collapses as the protein is vaporised by the electron probe.
There have been various discussions of this here on the forum before, and widely examined in the broader literature.
https://currentprotocols.onlinelibrary.wiley.com/doi/full/10.1002/cpmc.90
https://blake.bcm.edu/emanwiki/EMAN2/NegativeStain
And, of course, electron diffraction is a different case, e.g.:
Although electron crystallography is obviously not SPA.
There’s lots more out there in various EM texts and via a judicious search of PubMed, Web of Science or Google. 
3 Likes