October 19, 2012

New mtDNA age estimation paper claims post-LGM expansion dates for most relevant lineages

In Europe as in Africa and Native America.

You know that I'm very much refractory to molecular-clock-o-logy because it implies many a priori speculations and assumptions that probably make the results to be wrong and yet these results are presented as if they were the genetic equivalent of the proverbial "rocket science", when they are nothing but educated guesses, often contradicting each other.

Still I think that this paper is worth mentioning for two reasons:
  1. It gives age estimates that are clearly pre-Neolithic for most lineages, rebuking certain fanatic school that pretends to impose their clearly unfounded ideas of mass demographic replacement in the Neolithic and in a single direction, with a simple pattern. It is always good to have that kind of counter-arguments at hand.
  2. The paper is freely available (open access) in spite of being published by Nature (in an open access  dedicated magazine titled Scientific Reports).

Hong-Xang Zhieng et al., MtDNA analysis of global populations support that major population expansions began before Neolithic Time. Scientific Reports (Nature) 2012. Open access ··> LINK [doi:10.1038/srep00745]


Agriculture resulted in extensive population growths and human activities. However, whether major human expansions started after Neolithic Time still remained controversial. With the benefit of 1000 Genome Project, we were able to analyze a total of 910 samples from 11 populations in Africa, Europe and Americas. From these random samples, we identified the expansion lineages and reconstructed the historical demographic variations. In all the three continents, we found that most major lineage expansions (11 out of 15 star lineages in Africa, all autochthonous lineages in Europe and America) coalesced before the first appearance of agriculture. Furthermore, major population expansions were estimated after Last Glacial Maximum but before Neolithic Time, also corresponding to the result of major lineage expansions. Considering results in current and previous study, global mtDNA evidence showed that rising temperature after Last Glacial Maximum offered amiable environments and might be the most important factor for prehistorical human expansions.

Importantly, these age estimates must be considered as minimal ages (everything else assumed correct and equal) because the mutation rate used in at least some of the calibrations (Soares' rates) is clearly too fast, with even Dienekes (a former hardcore defender of very short chronologies and hyper-fast mutation rates) admitting to it as of late.

I quote from Dienekes:

Soares et al. were cautious, and they assumed an earlier Human-Chimp split than had been favored until then. However, a new paper by Langergraber et al. have used direct observation of the autosomal mutation rate and of ape generation lengths to argue for an even earlier Human-Chimp split: at least 7-8 million years ago, and as many as 13 million.

Which  is something I am quite glad that people has begun realizing because I was saying that more or less, in many different ways, since 2008 (refs.: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, etc.) And it was not just me because I was mostly reflecting what many scholars were finding here and there, publishing and seldom being listened to.

So these post-LGM dates are at best minimal dates, most probably requiring corrections that range between +15% and +85% (avg. +50%) per the Langergraber figures.

So for example, the authors estimate a 10-18 Ka window for the main expansions in Africa and Europe, that could well be a 15-27 Ka window in fact (max. 10-33 Ka.)

Figure 4: European expansion lineages from median-joining network
Blue, GBR; purple, FIN; yellow, CEU; cyan, TSI; green, IBS.

However mtDNA specifically has other major issues, absent in nuclear DNA (autosomes or Y chromosome): that the mutations happen so spaced in time for the whole mitochondrial DNA chain, which is very short, that the main factor is not the molecular tic-tac but population dynamics.

I should write more extensively on this matter but D. Pierron already did for me in 2011 (discussed here).

As I understand this problem of extremely unequal mtDNA lines, these irregularities are mostly caused by relatively large populations in which mutational change is effectively stopped by mere drift. As most women in the population experience no mutation in their mtDNA, the mutants have all the tickets to see their novel, derived, lineages drifted out by their relatives with a conservative one.

So in my understanding mtDNA "clock" must only be counted (if at all) from the root and not, as is usual, from the ends. This is a total game changer because counting from the root, H must be slightlty older than U, while counting from the ends that is impossible.

I never count mtDNA from the ends even if I know that most geneticists and aficionados do because I find such a procedure totally thoughtless, a mechanical imitation of a method that works for longer DNA chains which accumulate mutations in every generation, quite unlike mtDNA, which is much more stable.

En fin, the paper provides us with another opportunity to reconsider the highly speculative but incredibly popular field of molecular-clock-o-logy. It also provides some nice MJ networks of many haplogroups, which speak on their own, one of which (Europe), I used to illustrate this entry.


  1. I totally agree with your roughly ~+50% time correction: a lot of the European times for expansion in this paper are in the 15,000 to 20,000ya range, when we know there was no such thing. As I posted elsewhere, a factor of 1.5 pushes this into the Gravettian, when we know for sure there was pan-European expansion.

  2. "As I understand this problem of extremely unequal mtDNA lines, these irregularities are mostly caused by relatively large populations in which mutational change is effectively stopped by mere drift."

    It seems the precondition for this would be a community that is not expanding or growing. For a population that is expanding in territory, or growing in size as a results of new technologies, mtDNA lines with new mutations would be likely to survive and not be lost by drift. Given the trajectory of human migration and population growth, I wonder how often that condition would apply?

    1. Well, slow growth would be enough. If you have 100 fertile adult women (Ne) in generation no. 1 (G1) and 99 have haplogroup A and only one the newly mutated A1 (as must be the case with all new lineages), the chances may be 60% that A1 is extinct in G2, etc.

      If the population does not grow at very very fast levels all points to almost all novel mutations will vanish in few generations after birth or, the luckiest ones, remain as minority clades with no realistic chances of displacing the ancestral haplotype.

      Their best chances are actually a founder effect lottery: being part of a new colonization episode by chance (no DNA testing back then).

      Based on the mathematical/statistcal exercises I did some weeks ago with the help of a mathematics teacher, only for extremely low population levels (Ne=2, maybe a bit higher but certainly Ne<10) the tendency will be to not re-fixate. However there is always one remote chance that fixation happens in the novel lineage (I could not find how the exact chance might be estimated but the process can be simulated and that a novel lineage replaces the ancestral fixation seems to have an extremely low chance).

      A key issue is that mtDNA only seems to accumulate mutations every many many generations, unlike what happens in the much larger Y-DNA, where every new boy has a new haplotype almost for sure. Therefore with Y-DNA there's always change in each generation, so the re-fixation process is always happening once and again on novel haplotypes. Instead mtDNA is much more static and can be effectively frozen in a single haplotype (or a small group of them) by mere drift.

      With very fast growth as happens now, many new mutations accumulate surely also in mtDNA but fixating them is another matter and still only one of every 50 or 100 newborn girls (my rough estimate) will carry a novel mutation, each one different from the others, and every generation some 99% of those girls will still be having the same haplotype as their mothers. So, in a context of rapid growth like ours, a lot of new lineages have been created but how many will be consolidated and, critically, how many will manage to displace their ancestors' haplotype?

      It's a complex mathematical/statistical problem but it's clear that with some 99% of the descendants carrying the ancestral haplotype every generation, being everything else equal, the ancestral haplotype tends to stay and stay strong.

      The only "normal" way that a new haplotype seems able to create a niche is by means of small sized novel colonizations (founder effects) that later expand. So mostly when I see a node in the mtDNA tree, what I see is two small populations that part ways or, when there is a star-like node, what I see is a quick expansion into many small settlements or subpopulations from a single founder effect. The most dramatic of these are M and H, both of which show then a marked slowdown in the number of accumulated downstream mutations.


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).