Which population is most genetically distant from Africans?

By Razib Khan | August 18, 2010 1:58 pm

A comment below:

Razib, I don’t know much about genetics but is it true that these people of Melanesia are among the least related people (even more so than Europeans) to sub-saharan Africans genetically??

This is a common question. The typical scientifically curious intelligent person is generally aware that on the order of 100,000 years ago there was a movement of anatomically modern humans from Africa. They know that Africans have the most genetic variation of any human population, and that in fact Africa has more genetic variation than the rest of the world combined. It would stand to reason then that the further you are from Africa, the more genetically distant you are. Simply because of recent admixture of Sub-Saharan African ancestry in much of the Middle East there is some truth to this, but I think it misses the “big picture.”

To the best of my knowledge the current consensus on the origin and expansion of modern humans goes like so:

1) Anatomically modern humans emerge in Africa first ~200,000 years ago. This population is a sister lineage to the various Eurasian hominins, Neandertals, X-woman, etc.

2) Between 50,000 and 200,000 years ago a subset of the African population left Africa.

3) Sometime between the exit-from-Africa event and the present the anatomically modern humans replaced all other lineages (with some assimilation) and diversified.

My confidence in any specific aspect of the “orthodox census” is very high, though joint probability of the details is more modest. #3 for example had to be modified a bit recently because of the possible existence of Neandertal admixture in Eurasians. So back to my question, assuming this model, which population is most genetically distant from Africans? The answer is really none. Here are some figures from Xing et al., which gets at why the answer is “none of the above”:


Here’s the text for the figure:

Figure 3. Population relationships between the 40 populations. A) Neighbor-joining tree. Populations are color-coded based on their continental origins. The hypothetical ancestral population is shown. Bootstrap support values for most branches are larger than 95% (the bootstrap consensus tree is shown in Supp. Figure S1). B) Principal components analysis. First two principal components (PCs) are shown. Each individual is represented by one dot and the color label corresponding to their regional origin. The percentage of variance explained by each PC is shown on the axis. C) Individual grouping inferred by ADMIXTURE. Results from K = 4 and K = 12 are shown. Each individual’s genome is represented by a vertical bar composed of colored sections, where each section represents the proportion of an individual’s ancestry derived from one of the K ancestral populations. Individuals are arrayed horizontally and grouped by population as indicated.

The tree makes it clear: all non-Africans form their own independent branch from Africans. In the PCA you see that along the biggest component of variation in the genetic data the non-African groups are about the same distance from Africans. And in the ADMIXTURE analysis when you assume four ancestral populations, the Africans and non-Africans separate out cleanly excluding groups which a high likelihood of European or Arab admixture. Remember the part about how Africans have more genetic diversity than all non-Africans combined? That’s also part of the puzzle. In some ways all non-Africans can be thought of as a subset of the genetic variation of Africans. Those humans who reside outside of Africa are simply a diversified branch of Africans. From what I can tell the data is converging on the likelihood that there was only one migration out of Africa which resulted in the branches of non-African humanity. That means that those of us of non-African ancestry are all equally distant from the African root.

CATEGORIZED UNDER: Evolution, Genetics, Genomics
  • http://blogs.discovermagazine.com/loom Carl Zimmer

    So it’s the same thing as asking, “Which is closer to humans, chimpanzees or bonobos?”

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    i’ve been asked that too :-) (i gather you have?) in terms of phylogenetics i think it works well enough as long as you don’t look too closely at the details (africans would be humans since they didn’t diversify, right? but genetically humans are a lot less diverse than bonobos or chimps cuz we went through a recent bottleneck).

  • http://blogs.discovermagazine.com/loom Carl Zimmer

    Yup. Tree-thinking doesn’t come naturally.

  • Ian

    So non-Africans + Hema could be interpreted as sister groups, right? And since you can (usually) safely rotate any pair of branches, you could just as well put N. European next to Pygmy, and !Kung at the furthest end of the tree. So – “which population is most genetically distant from Africans?” could just as well be answered “Africans”.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    So – “which population is most genetically distant from Africans?” could just as well be answered “Africans”.

    ah. interesting. i guess i’ll need to dig up Fst numbers, as that paper didn’t have them listed. but yeah, i think that’s right. non-africans are a subset of northeast africans.

  • http://sep.stanford.edu/sep/jon/ Jon Claerbout

    One wonders whether PC3, PC4, … should be part of the story.
    I’m familiar with Principal Component analysis but unclear on how the genome is mapped into it. Is there an introductory article on that topic?

  • http://johnhawks.net/weblog John Hawks


    The charts don’t answer the question. Tree-thinking may not come naturally, but it doesn’t matter because a tree is a poor fit to human population distances. Reticulate thinking, people!

    If you want to know which population is most genetically distant from Africans, then you want to construct a distance matrix. There’s one hiding underneath that tree, with all the residuals thrown out.

  • Michael Watts

    The tree makes it clear that all non-African populations are equally separate from Africans *in time*. In a thought experiment where a subset of Africans migrate to the outside world (at the same time) and shortly thereafter diversify into two groups, one with a typical generation time of 20 years, and one with a typical generation time of 15 years, wouldn’t it make sense to say, after 100,000 years pass, that the group with the shorter generation time is evolutionarily further from the original Africans? They’ve been going for over 6500 generations to the other subgroup’s mere 5000. Would this result in a different-looking neighbor-joining tree? Is the whole idea flawed?

  • http://blogs.discovermagazine.com/gnxp Razib Khan



    john, well, i was gonna dig out Fst tables. need to have that handy.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    re: john hawks’ point, on page 20 are PCA’s derived from population wide Fst values


    interestingly the oceanic groups look closer to the sub-saharan populations along these two axes!

  • http://lyingeyes.blogspot.com ziel

    “interestingly the oceanic groups look closer to the sub-saharan populations along these two axes!”

    Phenotypically, that’s not that shocking, is it?

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    Phenotypically, that’s not that shocking, is it?

    phenotype has been a bad judge of this sort of thing (e.g., ainu, aboriginals, were assumed to be closer to europeans than to east asians, but that’s a false model). so you’re right, but we need to move away from phenotype when we have access to the genes.

    p.s. it’s a PCA representation of the Fst matrix, so i don’t put that much stock in the position of the oceanic groups in the HGDP

  • Pingback: Tweets that mention Which population is most genetically distant from Africans? | Gene Expression | Discover Magazine -- Topsy.com

  • DK

    which population is most genetically distant from Africans? The answer is really none.

    I don’t understand what you are saying here. It is absolutely given that there exists one. Only question is how you define the distance and which population it is. The most straightforward (and simpleminded) criterion would be Fst. So just look up Fsts. I don’t know the exact paper that lists them though.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    I don’t understand what you are saying here. It is absolutely given that there exists one. Only question is how you define the distance and which population it is. The most straightforward (and simpleminded) criterion would be Fst. So just look up Fsts. I don’t know the exact paper that lists them though.

    if all non-africans descend from one population of africans and there hasn’t been much gene flow between the two groups since the out-of-africa event (i don’t think there has, which is why i think john’s point about reticulation isn’t so important in this case) then they should all have about the same Fst from the ancestral population if you look across a lot of loci, right? i’ll look for the table, though i already wasted 15 minutes trying. seems like people have stopped including those tables now they got PCA and structure/frappe/admixture.

  • Melissa

    Couldn’t the answer be used to fuel hate groups who are looking for evidence to support their unrealistic beliefs? It’s an interesting answer for science and those who study genetic movement, but maybe there are some questions best left unasked.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    DK, i’m finding things which show Fst values varying as a function of distance from africa. the reason being serial bottlenecks and reductions in heterozygosity. so my answer might be wrong, i’m having a hard time finding tables in recent papers using SNP-chips.

    this leaves me with an interesting issue at i need to think about. my answer may have been more addressing the issue of relationship to last-common-ancestor, while extant variation is subject to evolutionary genetic parameters such as drift which have impacted peripheral groups of non-africans more than those fewer bottlenecks in (by this model new world populations are the most distantly related to africans). perhaps a follow up post.

  • Henry Harpending


    “but genetically humans are a lot less diverse than bonobos or chimps cuz we went through a recent bottleneck”

    This just isn’t true, it is an urban legend left over from the heady days of mtDNA. We are more diverse than gorillas and all but one chimp subspecies for which the data are poor. Orangutans are indeed more diverse than we are but many think that they are an amalgam of two different long separated populations, which would do it.


  • http://www.kinshipstudies.org German Dziebel

    “The typical scientifically curious intelligent person is generally aware that on the order of 100,000 years ago there was a movement of anatomically modern humans from Africa.”

    I’m all of the above (plus critical and cross-disciplinarily trained) but I don’t believe that data supports this scenario beyond doubt. It’s a possible model, but genetics – the crux of this model – has a lot of noise. For once, the out-of-Africa model is just a subset of the Multiregional-with-Admixture model, hence the Multiregional-with-Admixture model has greater posterior probability to be true (per Templeton). This doesn’t mean that the Multiregional model is the answer; it just means that both of them are overlapping and both of them may end up being wrong. Then, the fact that all non-Africans are equally removed from Africans (even when it comes to Neandertal “admixture”) may mean that Africans could have derived from any group of non-Africans, with selection or demographic pressures making Africans look distinct and diverse. Otherwise, it’s hard to believe that all non-Africans are so cleanly separated from Africans (see below on mtDNA). If we look at the biogeography of mtDNA, we’ll see that a) no African L lineages are found outside of Africa, b) only marginal/derived clades of non-African M and N macrohaplogroups are found in Africa, and c) most basal diversity in M and N macrohaplogroups is found in East and SE Asia (maybe India for M), thus creating a huge geographic gap between African and non-African clusters. Needless to say, if humans indeed migrated out of Africa along a southern route they would have left traces of their migration in the form of budding lineages. This is exactly what we see with M and N lineages but only going east-to-west into Europe and Africa, but not west-to-east (people may say at a later point, but still). Finally, other diversity factors, such as levels of linguistic diversity (grammatical and phylogenetic) show that Africa has 20 language families (and very few isolates), while places like America and Papua New Guinea (a geographic and often genetic polar opposite of Africa) has 70-140 families. This again speaks to differing demographic histories affecting Africans and non-Africans, with Africans possibly being a derived and expanding population, and non-Africans containing relics of unknown antiquity.

    As far as Fst values are concerned (also LD), Amerindians are on the opposite pole from Africans in most if not all systems. And they are geographically most removed from Africans.

  • http://entitledtoanopinion.wordpress.com TGGP

    I wasn’t aware there were more than two chimp subspecies. I thought it was just common chimp & bonobo. Razib might be referring to me above because I asked if (analogous to Africans vs non-Africans) we could examine their genetic diversity to determine which branches were oldest.

  • http://infoproc.blogspot.com steve hsu

    Here is an Fst table for you:


    It only has YRI (Nigerians) as the African sample, though. You can see that the lowest Fst relative to YRI are Romania and Spain at .14-ish, whereas the largest values are for China and Japan at just over .19. No Native American or Polynesian numbers given. In case you are curious the Fst between E. Asia and Western Europe is about .1 whereas, e.g., China and Japan are only separated by .007 and France-Germany by .0008. So the genetic “distance” (defined as Fst) across Eurasia isn’t that much smaller than the distance from southern Europe to Africa.

    You might find some of the figures at the link below of interest.


  • bioIgnoramus

    “maybe there are some questions best left unasked”: and who do you propose should identify them, melissa?

  • Sandgroper

    #16 You mean, let’s overcome hate and ignorance by stifling scientific enquiry?

    How would that work, exactly?

    The incidence of Type 2 Diabetes among Australian Aborigines is 3 to 4 times the incidence among the non-Aboriginal Australian population (and it’s already worryingly high among the whiteys and getting higher). Type 2 Diabetes causes 18% of all Aboriginal deaths. These numbers are suspected to be underestimates.

    That’s not because Aboriginal people eat and drink different stuff from the white folks and get less exercise, it’s because NOW they eat and drink the same stuff, and get the same amount of exercise (as in ‘not much’).

    Some sharp Aboriginal fellas are now advocating a return to traditional hunter-gatherer diet and lifestyle for Aboriginal people in order to improve health outcomes.

    How do you reckon them sharp fellas figured that out? By not thinking about genetic differences?

  • ghujhg

    I don’t know Melissa personally but there’s a good chance that, being a woman interested in the biological sciences, she’s far more conservative than men similarly inclined, by which I mean that she’s more likely to support the reigning regime (even when the reigning regime is radical).

    This is old hat, women tend to be more inclined to follow the leader than men are (even if only by a moderate amount and at their averages) and in his day and age the abolitionist—->60′s radicals have made their predilections the consensus. This isn’t to say that being inclined to trust the wisdom of your age is a worse idea than being inclined to oppose it but it is to explain why people “in the HBD-sphere” (how I hate that cultist language and the often rabid company) tend to be male by a ratio of about 10:1.

    Of course I’m more inclined to your view Bioignoramus, but I’m more attracted to Melissa.

  • onur

    “maybe there are some questions best left unasked”

    Welcome to the fascism of the modernies/liberals/leftists.

  • http://www.mazepath.com/uncleal/ Uncle Al

    KISS: Caltech.

  • Pingback: Which population is most genetically distant from Africans? Amerindians | Gene Expression | Discover Magazine

  • Pingback: Which population is most genetically distant from Africans? | Genetics News 2

  • http://entitledtoanopinion.wordpress.com TGGP

    Perhaps rather than embracing fascism as a form of governance she merely hopes that as a culture we come to realize that Ignorance is Strength.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar

Login to your Account

E-mail address:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »