Good work, Drew. I found it quite enthralling.

I think you can animate the problems of seeing what is going on just as you have described them here. Animate the clearing away of the obscuring molecules, identifying them in turn, to finally reveal the thing you want to show, then put them back afterwards as a reminder. Show a blur of brownian movement, then slow it down to a visible randomness then remove the random component to reveal the net effect. The Dynein walking can likewise can be shown erratically moving and net effect moving subsequently.

Reminders of these still non too visible realities may hover around as a smear of ghosting around the very solid "net effect" images for some of the time until the idea is grasped.

What you've just explained here about the basic realities of these molecular processes is just as interesting in its way to the novice as the more recent discoveries you focus on. Finding a visual (cinematic!) language for these unviewable things is the key and what you have made great progress in already.