Saturday, May 03, 2008
Effective Population Size of Our Ancestors in Africa
John Hawks has posted an interesting discussion of the effective population size of human ancestors. He concludes that Ne=34,000 during the late stone age (about one million years ago). According to Hawks, this means there were about 100,000 to 300,000 individuals spread throughout Africa at this time [Did humans face extinction 70,000 years ago?].
I don't understand the math, or the data. As a general rule, I'm skeptical of these calculations because so many papers seem to reach different conclusions. The really nice part of John's posting is that he tries to explain the assumptions and possible sources of error. It's worth a read just to get a feel for the kinds of things that are going on in population genetics.
There are no comments allowed on john hawks weblog. I'm sure if you post questions here he will answer.
Subscribe to:
Post Comments
(
Atom
)
7 comments :
Sure, thanks Larry. I'll check back if anybody has questions.
I want to thank John for this article. I was working my way through coalescent theory in order to give a much dumber account, but I am glad to see my intuitions confirmed by actual knowledge.
A couple of things. Could introgression of erectus mtDNA a la the multiregional model (at least as revised by Templeton) mess up the signal, and if so what should we expect?
Second, if Toba had caused a hominid bottleneck, shouldn't it have caused a similar one for other East African large mammals? Is there any evidence of that?
May be an obvious question: how do you distinguish between changes in population size and changes in mutation rate?
Speak for yourself, Larry. My ancestors come from india.
Peter wrote:
May be an obvious question: how do you distinguish between changes in population size and changes in mutation rate?
Unhappily, you can't. If we thought that the mutation rate was likely to change very much across the last 200,000 years, or on some mtDNA lineages more than others, we would have to give up the estimation of effective population size as hopeless.
Happily, we see no evidence for changes in the mutation rate across the last 200,000 years. Rates may have slowed somewhat on the human lineage compared to other primates, but not very much. What remains is the problem with accurately estimating the rate -- how many years does a mutation (on average) represent. The imprecision in this estimate is (at present) larger than any likely changes in the rate.
But there is a problem of saturation -- some sites are "hotspots" that have much higher chances of mutation than others, and over a long enough period of time, we cannot count these any more. Now that people have gone to whole-genome mtDNA sequencing, they can choose sites that are not hotspots, which cuts down on this source of error.
John Wilkins wrote:
A couple of things. Could introgression of erectus mtDNA a la the multiregional model (at least as revised by Templeton) mess up the signal, and if so what should we expect?
We know we don't have any recent human mtDNA that looks like known Neandertal mtDNA sequences, but what about African archaics, like H. erectus, or (if you use the term) H. rhodesiensis? Operationally, we have at least one mtDNA type from such a population, because the current mtDNA has ancestors. But the mtDNA variability that we see in today's human populations does not derive from the variability in the African population before 200,000 years ago. There has been too much inbreeding since then to retain the variability from that ancient population.
Second, if Toba had caused a hominid bottleneck, shouldn't it have caused a similar one for other East African large mammals? Is there any evidence of that?
It should have caused such bottlenecks. It apparently didn't. Ten years ago, it looked like there was evidence of a similar bottleneck in chimpanzees, as eastern chimpanzees (P. troglodytes schweinfurthii) have an mtDNA genealogy a lot like humans. But the other chimpanzees don't: today it looks like a recent colonization of East Africa by chimpanzees rather than an ancient bottleneck. Other mammals show no sign of such an event.
Also, with respect to Toba I should point out that there is good evidence from India of the Toba horizon, and at least one site has archaeology both directly above and below the horizon. It didn't even have a substantial local effect on humans, much less a global one.
Hi John Hawks,
I use discussions of effective population size to refute creationism, and I want to make sure that I have a solid one-sentence definition for "effective population size". Would you say that my following defintion is accurate: 'The "effective population size" is the hypothetical equivalent to the number of randomly breeding adults in a population.'
I guess I do not see how an effective population size could be 10 times the population size unless we are talking about major harems or high death toll before reproduction. Could you please explain more about the relationship of the population size to the effective population size?
Post a Comment