Mailbag: mtDNA mutation rates

Regarding frozen penguin mtDNA and mutation rates:

I read your comments on the ancient penguin DNA post on your blog, and the follow up where you mentioned that you don't think this study would push back the speciation date between humans and the other primates. But would this have an impact on the "recent out of Africa origin of modern humans theory"? Doesn't the penguin study call into question the estimated date of mitochondrial eve at 160ky BP and Y-chromosomal Adam at 60ky BP?

The basic question is how the “phylogenetic rate” relates to the “mutational rate”. The mutational rate can be measured by comparing mothers and daughters, this turns out to be very high. If these mutations were not under selection, each would be equally likely to reach fixation, and so the “phylogenetic rate” would equal the mutation rate.

But we know that it doesn’t, because chimpanzee mtDNAs are an awful lot more like us than they would be if they had diverged at the mutational rate. So, purifying selection must eliminate a large fraction of the new mutations that happen between mothers and daughters.

So far, so good. The question is whether purifying selection works in a few generations, or if it takes many thousands of years.

We know that human mtDNAs have an ancestor, but the time that we estimate for that ancestor depends on the pattern of selection – fast or slow, strong or weak? If purifying selection was very strong, then the only mutations that survive for very long must be neutral. Counting differences between populations will give us the same answer today as tomorrow, and the same differences would still be there after five million years – the phylogenetic rate, in other words.

If purifying selection was weak, then many of the mutations we see between populations today may be deleterious, ultimately bound for extinction. So counting differences today is an overestimate of how different the populations will be tomorrow. Something between the mutational and phylogenetic rates.

There has been quite a bit of literature on this topic in humans, so I don’t think the penguins add much except as a convenient example. I can’t say we really know the answer. There’s a lot of circular logic in the literature – people trying to estimate the rate based on assumptions about when humans reached India, for example, when those assumptions are themselves derived from mtDNA. It’s a mess.

The problem is worse for the Y chromosome, because the best estimates of the human most recent common ancestor are on the order of 50,000 years ago – clearly too young, but we don’t know how much too young.