Hominidae, Human, Mitochondrial DNA, Mitochondrial Eve, Mutation rate, Popular science, Radiocarbon dating, science
Most of the interesting recent events in human evolution probably happened longer ago than we had thought, according to Aylwyn Scally and Richard Durbin of the Sanger Institute. In an opinion piece published in Nature Review Genetics this week, they re-examine the story of how humans left Africa, taking into account new data from several recent whole-genome studies.
Scientists can determine the age of fossils quite accurately by using techniques like radiometric and incremental dating. Since fossils are uncommon, it’s fortunate that genetic techniques can also be used to establish the date of some events, such as when two populations split. There’s a problem, though: we can’t measure mutation rates. Instead, the mutation rate per year is estimated by comparing the genetic distance between two species with the age of fossils of their common ancestor. Of course, there are some caveats and qualifications, such as the effect of selection, but these can be taken into account; for example, genetic distance is usually measured by using stretches of DNA which we think aren’t subject to selection. This technique enabled scientists to estimate a mutation rate for humans by using the genetic distance between humans and orang-utans and the age of their common ancestor; a similar mutation rate was found using the divergence between humans and macaques and fossil evidence for the split between apes and Old World monkeys. This mutation rate was then used to calculate the timing of various events in prehistory, such as when some modern humans migrated out of Africa or when our lineage diverged from Neanderthals.
With the advent of fancy, high-throughput next-gen sequencers, though, it’s become possible to directly measure the mutation rate in humans. Since we can now sequence entire genomes with relative ease, measuring the mutation rate is as straightforward as counting the number of differences between the sequence of a child and its two parents: any changes in the child that aren’t present in either parents are mutations. The average number of mutations from lots of parent-child trios is the mutation rate per generation. This is how the mutation rate was determined in the widely reported recent study about older fathers passing on more mutations to their children; other studies have found a similar mutation rate. In order to compare these rates with previous estimates, Scally & Durbin divide them by the average generation time to convert the rate per generation into a rate per year. Using a generation time between that of modern humans (30 years) and chimpanzees and gorillas (20 years) gives a mutation rate around half the value found by divergence estimates. In order to get the same value, generations would have to be an unlikely 10-15 years long — half the value in present-day humans and less than any living great ape. The researchers suggest that one explanation for this discrepancy is that the mutation rate has gotten slower in great apes since their divergence from other primates; this idea, called the hominoid slowdown, has some support from other research.
Halving the mutation rate of our ancestors means that we have to double the dates of many of the interesting events in human prehistory. I started my last post by saying that modern humans spread from Africa around 60,000 years ago; now, just a week later, that figure has been revised to somewhere between 90,000 and 130,000 years ago. The new estimate actually fits well with archaeological evidence which suggests that modern humans might have been present in the Middle East earlier than 60,000 years ago (the Skhul/Qafzeh people). These had previously been thought to be the remains of a temporary excursion out of Africa; based on the new date, however, they may in fact represent later stages of the migration itself. The revised date also pushes back the age of the split between European and Asian populations to 40,000-80,000 years ago. While this fits better with the archaeological evidence, it leaves a period of about 50,000 years between the migration out of Africa and the colonisation of Eurasia. The authors suggest that the migration may have happened in stages with the intermediate population initially confined to the Middle East, providing a good opportunity for interbreeding with Neanderthals.
The new mutation rate has also helped resolve questions about when modern humans and Neanderthals diverged. While previous studies based on mitochondrial DNA have estimated a date around 500,000-600,000 years ago, studies using nuclear DNA give a much more recent estimate. Since the new mutation rate was estimated using nuclear DNA, it brings the results of these different studies into agreement, giving an estimate of 400,000-500,000 years ago.
The lower mutation rate also changes the date of the split in modern humans, between the Khoe-san people and others. The new estimate of 250-300,000 years ago is somewhat problematic, since it’s older than the age of the
most recent common ancestor of all modern humans most recent common mitochondrial ancestor of all living humans, which has been dated at 125,000-250,000 years ago using mitochondrial DNA. [Thanks for the correction, Jack!] This may result from some technical issues with using mitochondrial DNA to estimate when populations diverged, but it might also reflect a more complicated population structure in ancient Africa.
We may find ourselves continually rewriting our history of ourselves over the next few years as genomics technologies improve and more data become available. While the broad strokes of the story will probably stay the same, we should bear in mind that a whole new hominid — the Denisovans — was discovered less than five years ago. Amazingly, pretty much everything we know about the Denisovans comes from genomics, since we only have a few fragments of bone. Earlier today I heard about some research which used the genome of the 5,300 year old Iceman Ötzi to study migrations in ancient Europe. It’s an exciting time in the field of human evolution and there will certainly be loads of interesting new findings coming. Hopefully closer collaboration between geneticists, anthropologists and archaeologists will help us unravel the story of how our ancestors spread across the planet.
Scally A, & Durbin R (2012). Revising the human mutation rate: implications for understanding human evolution. Nature reviews. Genetics PMID: 22965354
Razib Khan (@razibkhan) said:
i think their piece was more equivocal than that. though they think that their piece needed to be written since there’s a high probability that mutation is lower….
To be honest, I didn’t sense much equivocation when I was reading the paper. The revised dates I discuss in this post are what the authors call the “four key points” in their conclusion. While they are tentative about some of their conclusions/suggestions (eg, the question of ancestral population structure, the intermediate migration or the timing of the split with Neanderthals), they seem quite confident that the current interpretation is wrong since it’s based on too high a mutation rate.
Aylwyn Scally said:
Hi, just saw this – thanks for blogging about our research. I would say that is a fair summary of our view.
Glad to hear I didn’t misrepresent you. I enjoyed the article and your talk in Helsinki!
Interesting research, thank you for your post.
“The new estimate of 250-300,000 years ago is somewhat problematic, since it’s older than the age of the most recent common ancestor of all modern humans, which has been dated at 125,000-250,000 years ago using mitochondrial DNA.”
First up, the most recent common ancestor is NOT mitochondrial eve! She is the most recent common female ancestor traced exclusively down the female line. The most recent common ancestor is much, much more recent than that (maybe as recent as 3-5 thousand years ago, IIRC).
Secondly, it would be no surprise that the mitochondrial eve is more recent than distinct splits in population structure. Given that humans are all capable of interbreeding it follows that some small level of gene flow will continue to have occurred between populations; such small level gene flow permits genetic differentiation between populations but is still likely to eventually unify mitochondrial lines unless the transfer is exclusively male.
Thanks for your comment!
“First up, the most recent common ancestor is NOT mitochondrial eve! She is the most recent common female ancestor traced exclusively down the female line. The most recent common ancestor is much, much more recent than that (maybe as recent as 3-5 thousand years ago, IIRC).”
You’re absolutely right. I meant most recent common mitochondrial ancestor, but I stumbled over the sentence because it was getting late. My mistake — thanks for calling me out. 🙂 [I’ve corrected the mistake Jack pointed out.]
I agree that the mismatch between the mitochondrial tree and the population divergence isn’t really an issue. There are several possible explanations, including the gene flow scenario you mentioned, as well as the smaller effective population size for mtDNA, the possibility of recurrent mutations and the fact that mtDNA is a single locus. I glossed over those in the main text (“technical issues” and “population structure”) because I’m trying to keep my posts relatively light, but thanks for bringing them up!
Thanks for the speedy correction 🙂