[jsnell]

If only half the population in Europe shared a most common ancestor 4000 years back, that would be absolutely shocking. But this article is misleading, and that's not what the research says.

It's actually not about ancestry in general, but purely about patrilinear ancestry [0], which doesn't have the same kind of fan-out. If you go 30 generations back, you might have a billion ancestors, but only 30 patrilinear ones.

[0] That's the measurement they have to make. We have technology to measure patrilinear ancestry through the Y chromosome, and matrilinear ancestry through the mitochondrial DNA. But there is no way of doing the same for ancestry in general.

[maxander]

> But there is no way of doing the same for ancestry in general.

Uh, sure there is- apply the same techniques to non-Y, non-mitochondrial chromosomes.

[jsnell]

I don't think you can apply the same techniques, since they depend on a special property of the Y-chromosome and mitochondrial DNA: they are not affected by sexual reproduction, only by mutations.

[maxander]

Huh, you're right- studies on the altogether most recent human common ancestor are all based on pure statistics. I'd thought that genetic recombination could be accounted for, since its not as though genetic information (and the mutations therein) are being lost in the process, but apparently no one's found a way to do this yet. Interesting... ! :)