Population structure, concrete and ineffable

Before there was Structure there was just structure. By this, I mean that population substructure has always been. The question is how we as humans shall characterize and visualize it in a manner which imparts some measure of wisdom and enlightenment. A simple fashion in which we can assess population substructure is to visualize the genetic distances across individuals or populations on a two dimensional plot. Another way which is quite popular is to represent the distance on a neighbor joining tree, as on the left. As you can see this is not always satisfying: dense trees with too many tips are often almost impossible to interpret beyond the most trivial inferences (though there is an aesthetic beauty in their feathery topology!). And where graphical representations such as neighbor-joining trees and MDS plots remove too much relevant information, cluttered Fst matrices have the opposite problem. All the distance data is there in its glorious specific detail, but there's very little Gestalt comprehension.