Yeah. One might for example reduce reinforcement of the big-empty-cell misconception by briefly showing more realistically dense packing, eg [1], before fading out most of it to what can be easily rendered and seen. But that would be less "pretty". Prioritizing "pretty" over learning outcomes... is perhaps a suboptimal for education content.

> better

But still painful. Consider those quiet molecules in proteins, compared with surrounding motion. A metal nanoparticle might be that rigid, but not a protein.

One widespread issue with educational graphics, is mixing aspects done with great care for correctness, with aspects that are artistic license and utter bogosity. Where the student or viewer has no idea which aspects are which. "Just take away the learning objectives, and forget the rest" doesn't happen. More like "you are now unsalvageably soaked in a stew of misconceptions, toxic to transferable understanding and intuition - too bad, so sad".

So in what ways can samplings of a protein's configuration space be shown? And how can the surround and dynamics be shown, to avoid misrepresenting that sampling by implication?

It can be fun to picture what better might look like. After an expertise-and-resource intensive iterative process of "ok, what misconceptions will this cause? What can we show to inoculate against them? Repeat...". Perhaps implausibly intensive. I don't know of any group with that focus.

[1] https://www.flickr.com/photos/argonne/8592248739