Evidence That Two Main Bottleneck Events Shaped Modern Human Genetic Diversity – Proc R Soc B FirstCite

The subject of bottlenecks in ancient human populations is visited once again, as Amos and Hoffman propose to have found evidence for two such events, one as humans migrated out of Africa and later when a migration event into Pleistocene America occurred across the Bering Strait.

Here’s the abstract of the paper which is freely accessible:

There is a strong consensus that modern humans originated in Africa and moved out to colonize the world approximately 50 000 years ago. During the process of expansion, variability was lost, creating a linear gradient of decreasing diversity with increasing distance from Africa. However, the exact way in which this loss occurred remains somewhat unclear: did it involve one, a few or a continuous series of population bottlenecks? We addressed this by analysing a large published dataset of 783 microsatellite loci genotyped in 53 worldwide populations, using the program ‘Bottleneck’.

Immediately following a sharp population decline, rare alleles are lost faster than heterozygosity, creating a transient excess of heterozygosity relative to allele number, a feature that is used by Bottleneck to infer historical events. We find evidence of two primary events, one ‘out of Africa’ and one placed around the Bering Strait, where an ancient land bridge allowed passage into the Americas. These findings agree well with the regions of the world where the largest founder events might have been expected, but contrast with the apparently smooth gradient of variability that is revealed when current heterozygosity is plotted against distance from Africa.

The researchers suggest that their more detailed approach to investigating the data allows for a more complex picture to emerge, which in the process threw up some unexpected findings, as revealed towards the end of the paper:

Despite these complications, a rather consistent pattern emerges, with evidence of a bottleneck being strongest in the Middle East and in the easternmost East Asian/northernmost American populations. These two locations are as one might expect, but there are two additional features that are less obvious. First, the African populations, although at most loci having low t-values, do provide quite strong and consistent evidence of a bottleneck at the lowest variability loci. As discussed, this may reflect an observation bias in which loci with very low variability in Africa are unusual for some reason other than demography.

An alternative explanation is that these loci still retain the signal of an even more ancient, within-Africa event. This would be consistent with the notion that locus variability is inversely related to the antiquity of the bottleneck signal that is best retained and offers an intriguing hypothesis for future studies. The second feature is the pronounced dip in t-value between Europe/central southern Asia and East Asia. This may simply reflect a null signal between two bottlenecks, but might alternatively indicate some other demographic event such as a period of stasis and population expansion. Again, further work is desirable.


Evidence that two main bottleneck events shaped modern human genetic diversity by  W. Amos and  J. I. Hoffman, 2009 – Published online before print October 7, 2009, doi: 10.1098/rspb.2009.1473

22 thoughts on “Evidence That Two Main Bottleneck Events Shaped Modern Human Genetic Diversity – Proc R Soc B FirstCite

  1. Thanks for posting this, Tim.

    I’ve been taking a look at the paper and the corresponding data (Rosenberg 2003 & 2005) and I’m somewhat puzzled when I try to check individual populations.

    I did that because I understand that raw distances to Africa could be meainingless, in particular for West Asia, in reflecting population history. This applies both with the mainstream “coastal migration” model and the sometimes suggested alternative path though Central Asia.

    While I have not been able to find a t-value list for each of the populations, the element that best approaches is figure 4 (populations are not named but at least defined by regional adscription). And there you see clearly that the lowest “bottleneck values” (t-values) are not just for some African populations but also for some CSA and (oddly enough) for most American ones. Even one European population (which one?) is also at the bottom t-value levels.

    When you take the regional clusters though, the picture is more as the authors claim, with the overall Africans apparently holding the lowest t-values, closely followed by CSA (Central/South Asians, which in Rosenberg’s data are mostly Pakistanis and Afghans, with only the Uyghur sample being truly “Central Asian”). This would be consistent with the coastal migration model.

    However, after CSA, and quite close to them, are Europeans and Oceanians, followed closely by East Asians. Europeans here are quite odd considering the known prehistory and the low haploid diversity, and also if we compare with the very limited Middle East sample (Druzes, Palestinian Bedouins and Mozabites, all very peculiar populations, and Palestinians), which show high bottleneck signals, both as individual populations and as a group.

    I have ignored Native Americans in this regional analysis because I could not just by eye define a cluster center, but with 3/4 samples at the lowest t-values of the whole graph, they would look rather not bottlenecked.

    Nevertheless, the key analysis for a Native American bottleneck comes not from t-values but from what they call the raw data. Raw data that I could not see listed anywhere (though maybe I have not looked deep enough). So these impressions I’ve got from t-values may be misleading.

    1. Luis – Thanks for your comment.

      As far as I can tell the raw data to which they refer is mentioned under ‘Materials and Methods’ in which they include the following link:

      “Data for 783 microsatellites genotyped in 53 worldwide populations were downloaded from http://rosenberglab.bioinformatics.med.umich.edu/diversity.html (Rosenberg et al. 2003, 2005). Bottleneck analysis was conducted using Bottleneck v. 1.2.02 (http://www1.montpellier.inra.fr/URLB/bottleneck/pub.html) (Luikart et al. 1998). We explored a range of mutation models, from the strict SMM through three models with varying proportions of jump mutations (two-phase models, TPMs), to the IAM. For the TPM we used the default variance of 30 and three different proportions of jump mutations: 10, 5 and 2 per cent.”

      with the specific link being:


  2. Thanks a lot, Tim. I could not find the specific repository.

    Not that I think I’ll be able to do much with that mass of polymorphisms (apt only for computerized processing), but it’s always good to know it’s freely available to the public.

  3. It’s good to see you, Luis, coming to your senses. The idea of a bottleneck on the way to America originally emerged as an awkward way to explain high frequencies of rare Asian markers in Amerindian populations. The bottleneck idea was quickly debunked by Ward et al. (Ward, R. H., Frazier, B. L., Dew-Jager, K., & Pääbo, S. (1991). Extensive mitochondrial diversity within a single American Indian tribe. Proceedings of the National Academy of Sciences, 88, 8720–8724) and Chakraborty and Weiss (Chakraborty, R., & Weiss, K. M. (1991). Genetic variation of the mitochondrial DNA genome in American Indians is at mutation drift equilibrium. American Journal of Physical Anthropology, 86, 497–506). There’re only a few tribes in the Americas (Cayapa in Ecuador being one of them) that showed signs of a bottleneck. For the most part, Amerindian populations are unbottlenecked. They show relics of Pleistocene population structure, whereas Africa is a continent that absorbed a few migrationary waves between 45,000 and 20,000 YBP that brought in several bottlenecked populations that quickly started expanding in size and exchanging genes and made this region by 1492 the most genetically diverse. Eurasian populations haven’t grown in size as much as African populations but, with the exception of Papua New Guinea and Australia, experienced strong population growth that resulted in many lineages not observed in Amerindian populations. After 1492 the pendulum swung back and brought subsets of European, Asian and African populations into the Americas and now made this region the most genetically diverse.

  4. I’m not “coming to my senses” in the way you’d like, German. As I haven’t got the software myself (seems freely available but I don’t want to bother), I will have to take the authors results on the raw data at face value. I always prefer not to but in this case I would have to work too much to make a critical assessment.

    Now, I don’t like the use of the term “bottleneck” anyhow because what they mean is in fact a founder effect. A bottleneck is actually a massive death of 90% or more (often more than 99%) of the population, causing a massive reduction in genetic diversity. Bottlenecks and founder effects have similar effects but are not the same. In fact massive bottlenecks is what I understand you claim for your “Out of America” hypothesis to have happened in America and Eurasia but I just make no sense of it all. As we have discussed already a zillion times, neither archaeology nor genetics nor any other discipline except your quite peculiar reading of kinship structures, where complexity precedes simplicity, support that. You are stubborn but stubbornness alone won’t bring you close to truth. In fact science has a lot of being flexible and being able to change opinion when the evidence suggests it’s the right thing to do.

  5. There’s nothing in archaeology that unequivocally supports out of Africa. What archaeology tells us is that there was population growth occurring simultaneously in Africa and Europe from 45-40,000 YBP on, which is AFTER the proposed expansion out of Africa. European Upper Paleolithic documents not a rapid replacement but rather continuity. Same for East Asia.There’s no trace of an out of Africa migration in Asia and Australasia but rather a slow accrual of modern human behavioral signatures up until 10,000 YBP.

    Genetically, as indeed discussed ad finitum, African lineages are not found in modern populations outside of Africa. The earliest offshoots of M and N macrohaplogroups aren’t found in Africa. Hence, we have the absence of 1) African-specific lineages outside of Africa as well as 2) potentially ancient non-African-lineages within Africa.

    Neither do we find African mtDNA L lineages in 28,000 year-old Cro-Magnon remains. mtDNA L lineages are not even found in Taforalt remains in North Africa (12,000 YBP). There’s therefore no direct evidence for the antiquity of African lineages.
    What we would expect, if out of Africa was good science, is to find L lineages in modern populations outside of Africa at low frequencies (especially, along the proposed coastal route into Australia) and L lineages in ancient remains outside of Africa at higher frequencies. This is demonstrably not the case.

    I do propose a series of bottlenecks for the populations entering Africa. This is more than natural assuming that small foraging demes had to traverse vast terrains, unknown ecological niches and climatic zones separating America from Africa. Once they got into Africa, they expanded in size and mixed with each other through recurrent gene flow.

    Another possibility is that mutation rate is higher in African lineages, but even a neutral demographic model can account for the late origin of African genetic diversity.

    As for your (mis)understanding of kinship evidence, it’s not that complexity precedes simplicity (you’re confusing Victor Grauer’s musicological logic in which polyphony universally deteriorates into monophony) but that a set of very distinctive, intricate but symmetrical patterns of kinship terminologies can be shown to exist only outside of Africa, while their traces (plus a couple of clearly derived complex forms) are found in Europe and Africa. Out of America explains this pattern very neatly. It’s possible that the frequencies of ancestral forms of kinship and social organization as well as the levels of linguistic diversity are high outside of the homeland and low in the homeland just through some kind of magical “mutational wave” effect, but I still wouldn’t outright dismiss a more parsimonious explanation that I advocate for. Especially since there’re problems with out-of-Africa in a few key areas of proof (low linguistic diversity in Africa, no linguistic isolates, no direct evidence for the antiquity of African genetic lineages, large effective population size in Africa vs. low effective population size attributed to Mid-to-Late Pleistocene humans, an archaeological signal of population growth in Africa and Europe, the association of “anatomically modern humans” with Mousterian technologies, etc).

    Luis: “In fact science has a lot of being flexible and being able to change opinion when the evidence suggests it’s the right thing to do.”

    German: Go ahead and be flexible. I’ve got the evidence.

  6. “What we would expect, if out of Africa was good science, is to find L lineages in modern populations outside of Africa at low frequencies”.

    Actually we do, but not at low frequencies. Virtually all non-African lineages belong to just two L3 lineages. Now, that’s difficult to account for other than that these two lineaged came out of Africa. Of course we’re only considering the mtDNA here but the same holds for Y-chromosome. But even that doesn’t necessarily mean all human genes came out with those haplogroups, or even that the original mtDNA and Y-chromosome haplogroups emerged from Africa together.

    I’m quite prepared to accept movement from, say, Alaska back into Asia but to accomodate the whole of modern humanity to a movement out of America is extremely difficult to reconcile with ALL the evidence. It’s possible only if you selectively ignore portions of it.

  7. Terry,

    Your argument is a circular one: if the trees were constructed in such a way that M and N were derived from L3, then M and N ARE L3. Hence, all non-African diversity is African diversity. Would you say that all lineages are L0 and we are all Bushmen? If you apply the same logic to geography, you could say that America is a tip of Africa, hence humans haven’t really left Africa. This is not the best way to prove things. The current tree topologies need to be proven; instead you assume that they ARE the proof. Ancient DNA is one way to prove or disprove things but you seem to absolve geneticists from the need to prove their phylogenies. By doing so you admit a logical flaw by assuming that whatever is found in ancient remains outside of Africa is derivative of modern African populations. However there’re good reasons to believe that time flows the opposite way.

  8. There’s nothing in archaeology that unequivocally supports out of Africa.

    There is nothing in kinship that unequivocally supports out of America, German. Absolutely nothing! You’re beating a dead horse and I’m bored of finding the corpse at every other corner. It’s beginning to stink.

    1. What do you know about kinship studies, Luis? It’s the oldest tradition in anthropology operating with the largest databases of human populations and with decades of experience building evolutionary models of non-biological traits. It’s a critical piece of the puzzle capable of connecting linguistics and population genetics through a model predicting a pattern of correspondence between linguistic features and demographic and social structure, which in turn influence the patterns of molecular variation.

      In any case, kinship systems only suggested an alternative to out of Africa which is not Multiregional. I base all my inferences on the convergence of data points across all sciences from genetics through craniology and dentition to linguistics and history of science. Every discipline should be kept within its own limits of effectiveness and the truth comes from aligning the most secure aspects of the data supplanted by one discipline with the most secure aspects of the data supplanted by another discipline.

      Out of Africa has too many problems to not look for alternatives. You’re welcome to suggest other original ways to think about the data but your style of regurgitating selections from out of Africa favorable publications doesn’t add much to the already existing perspective that, after spending almost 10 years at Stanford, I know all too well.

      After reading your comment on the Amos and Hofman paper, I thought you began seeing the same pattern as myself but apparently I was wrong. See, I’m eager to admit my own mistakes when the evidence is there.

  9. What do you know about kinship studies, Luis?

    Enough. I know also, that like other “humanistic” disciplines (linguistics, anthropology in general) is prone to subjective interpretation. Slippery, largely speculative, science that must be taken with great caution, where most of what is written (apart of the field data) is just educated opinions.

    I will also tell you another thing (and please Tim or Kambiz, feel free to cut this rather off topic discussion whenever you think convenient): American Natives are the only continental population worldwide where tropical peoples are not black (dark brown to black skin color). This is a well known fact and is because they lost their tropical pigmentation when migrating to Siberia and Beringia and have not got evolutionary time to regain it. This alone is enough evidence to discard your hypothesis, though of course there is much more.

  10. Luis,

    I thought we’ve already discussed this issue at some point: the skin color of the Khoisans, which both Y chromosome and mtDNA studies nominated to be the oldest human population, is controlled by a derived allele. It’s lighter than that of other blacks and is often likened to the skin color of Mongoloids. E.g., “The lightly pigmented hunter–gatherer San population of Southern Africa is exceptional in having a high frequency of the derived allele relative to geographically proximate and more darkly pigmented African populations” (Norton et al. 2007, 719).

    So, we don’t really know what skin color the earliest humans had. Maybe it was intermediate exactly like that of American Indians. Your theory of American Indians losing tropical pigmentation is contradicted by the Amos and Hofman paper under discussion, which, as you yourself pointed out, demonstrated that Amerindians are the least bottlenecked among the human populations in their sample.

    The only way to control for subjective biases is to painstakingly weigh different options and to take all disciplines and all data into consideration. We’re in the business of human origins. Since when have “humanistic” and “anthropological” become curse words?

  11. I don’t know if to love or to hate how you manage to divert the issue. I am talking of skin color of (some, many) tropical Africans, Asians and Oceanians in contrast to that of Native Americans. Khoisans live south of the tropic of Capricorn mostly, so it does hardly apply.

    It is evidence that Native Americans had not enough evolutive time to develope anything of the sort, because they arrived only in “recently”, maybe some 15,000 years ago. Instead elsewhere they had plenty of time, maybe two million years or more if we count from H. erectus.

  12. Luis – Apart from the fact I don’t think it’s within my editorial remit to terminate threads, I’m of the firm belief that correspondents should be able to comment as they see fit – just as it’s the readers’ choice to follow or ignore such threads as they wish.

    From a personal point of view, I’d prefer people who clearly have a great deal to say on whatever topic, to write their own blog posts/articles in which all their thoughts are laid out clearly and coherently, rather than engage in the kinds of diatribe that all too often end up with insults traded and credibility questioned – a zero sum game that is of benefit to few, if anyone at all.

    Alternatively, where instances of extended threads that depart from the original post become too commonplace, I’d suggest that there are innumerable fora elsewhere online which might be more suitable venues for such discussions to take place.

  13. The paper under discussion is about the two bottleneck idea and not about skin color.

    In any case, I expounded on an alternative, out of-America, view of ancient human dispersals in The Genius of Kinship: The Phenomenon of Human Kinship and the Global Diversity of Kinship Terminologies (Cambria Press, 2007) as well as here (https://anthropology.net/2008/05/12/the-genius-of-kinship-human-kinship-systems-and-the-search-for-human-origins/).

    The Amos and Hoffman paper offered some evidence that is consistent with out-of-America and contradicts out-of-Africa.

  14. yes the comments here are disenchanting, arethere really people who argue unarcheological idea’s like out of canada? amazing. and indeed what does it have to do with skincolour? unfortunately i don’t understand the research, too vague, an allel here or there, perhaps people doing such research would be helped to know that in the past 3 century’s most of the african populations, and thus genes have been driven south, and it explains some of the relative uniqueness of the khoi dna. (that you could better compare to zulu and xhosa eg.) also i still have to see a credible explanation for a bottle neck event in the levant, if anywhere imo that would have been in egypt, so in africa, also.. i may be misinterpreten, but any genetical indication for the whole of humankind, must at some point have been defining in our stature, thus.. the rare chance of a paralel mutation left out of consideration, you might want to define many more bottleneck events, especially in the early history of hominids, i think afar shows how that went.

  15. pls i will like to ask a question, i am a student in one of the university in Nigeria ( Part time )
    Q1, Anthropology has a pride of place in the social science because of its holistic multifaceted approach to the study of human beings. Expatiate
    Q2, Anthropology methods and theories are relevant and adequate in the study contemporary. Pls Elaborate.

Comments are closed.

A WordPress.com Website.

Up ↑

%d bloggers like this: