Discover more from Razib Khan's Unsupervised Learning
Out-of-Africa's midlife crisis
on bottlenecks, crashes and what diversity really looks like
Life comes at you fast
Maybe this is just a millionth testament to our species’ freakishly strong predisposition to pick up any possible tool and start bashing things with it. But it kind of beggars belief how quickly we come to consider something almost impossibly miraculous... commonplace. In the same century where life-and-death court proceedings could hinge on an instrument as blunt as blood type, today geneticists just routinely peer inside the entire human genome, meticulously reconstructing the story of our species, letter by letter.
Now that Prometheus has basically granted every lowly lab tech superhuman powers, it’s like drinking from a data firehose if you love ancient DNA. And the more we know, the clearer it looks that genetically, all the humans on our planet group into basically three genetic types. You could think of them as 1. the very-diverse, 2. the not-very-diverse 3. the very-not-diverse (and if we’re being thorough: 4. a recent hybrid of 2 and 3.)
Do you know which you are? You don’t have to tell me anything else about you; by the very fact that you are reading this I can almost guarantee you are in group 2 or 3. Just based on the numbers, I’ll put my money on 3. Six to one odds. (And if you’re in group 1 and haven’t been genotyped but are interested in human genetics, DM me! I can help you get a spit kit. You’re genetically very unique!)
Here is how the groups break down:
Scarcely half a million of us are very-diverse.
1.14 billion of us are not-very-diverse.
6.42 billion of us are very-not-diverse.
Alone on our planet today, those maybe half a million very-diverse souls hint at our species’ one-time amazing levels of genetic diversity. In our DNA, we all contain multitudes. But once, we all contained mega-multitudes. Only the very-diverse retain much of it today.
The genes must be crazy: genetically very-diverse humans
The very-diverse are hunter-gatherer populations who live in clusters scattered across Central and Southern Africa, in Namibia, Botswana, the Congo and a dozen other nations. They include the Khoi, the San, the Mbuti, the Mbenga, the Twa and the Hadza. Inside their DNA, they carry such a dazzling diversity of single-nucleotide-polymorphisms (SNPS), that even two San from different groups both living in Namibia’s Northern Kalahari desert, and speaking click languages from the same family, are more genetically distinct from one another, by a solid 20%, than a person from Stockholm is from a person from Shanghai. That is, they average a rate of 1.2 nucleotide differences per kilobase (1000 SNPs), where a Northern European and Chinese person differ from each other at a rate of only one difference per kilobase. And the San in this example were both from the Kalahari; imagine comparing samples from peoples who live thousands of miles apart!
The DNA results are in. You are.... NOT genetically diverse (all 7.7 billion of you)
Why are the vast majority of us not-very-diverse and very-not-diverse? And which is which? Genetically, all the rest of modern humanity is one of two similar flavors (or a swirl of the two) with lessened genetic diversity, because along the path to today, we lost most of our diversity. How? Severe population contraction. Inbreeding, if you must. As far as we can tell, all of humanity, at some point in the past 60,000-120,000 years was forced through one or more funnels. And most of our glorious diversity of potential ancestors didn’t survive the successive culls.
In the case of the very-not-diverse (fully 6.42 billion-ish of us), our models suggest that life on earth threw a funnel across their path, and only 1,000-10,000 lucky humans made it through the impossibly narrow neck. Our best (always provisional) models suggest they dwindled from an earlier population maybe 10 times bigger. So this round of cuts diminished the breeding pool by 90%. Think about that. Only 1,000 to 10,000 total humans could survive whatever those harsh conditions were. On a planet with 7.7 billion humans today, that small a group is a rounding error, the student body of a large high school or a small college, an isolated mountain village that could disappear tomorrow without anyone noticing. How close did we come to losing the ancestors of every single one of 6.42 billion of us alive today? Well, we came within 1000-10,000 people of it.
Go forth and multiply
Those 1,000-10,000 human beings who made it through their ordeal, smuggled out in their nuclei all the genetic diversity 6.42 billion very-not-diverse humans among us today would have to draw on ever after. Take a native each from say Santa Fe, Stockholm, Shanghai, Ulaanbaatar in Mongolia, Uluru in the Australian outback and a Sentinelese person from the Andaman Islands and you behold a little of the amazing superficial variety of the human race. They and their relatives have peopled almost every corner of the earth. They speak a riot of different languages and they look nothing like one another. And yet, aside from a dash of Denisovan here and a trace of Neanderthal there, as far as we can tell, they all trace the entirety of their ancestry back to a single founder event about 60,000 years ago. An event when just a tiny subsample of 1,000-10,000 humans of that day passed through a brutal, extended bottleneck. Whether it was in Africa or just after leaving the continent, they burst out of that dire strait and re-peopled the globe.
But bottlenecks are not rare. They can and do happen to any population, any species. Have you met a bulldog? In the case of humans, we can already tell that the ancestors of all of our extant groups today had to squeeze through bottlenecks somewhere between 60,000 and 120,000 years ago. And not just the very-not-diverse. But as you probably already guessed, the very-not-diverse got hit the hardest, the very-diverse the least and the not-very-diverse somewhere in between.
It takes a small city
We’ve covered the student-body’s-worth of genetic diversity that improbably went on to settle five continents and the habitable islands strung between. But that leaves all of Africa. How genetically diverse are most Africans? Scattered in their midst are tiny clusters of the planet’s most genetically diverse humans. But genetically, all other Sub-Saharan Africans more closely resemble their bottle-neck-surviving brethren who left for other continents than they do their own very-diverse geographical neighbors.
Save for that isolated half-million very-diverse hunter-gatherer minority, all Sub-Saharan Africans constitute the genetically in-between group of 1.14 billion not-very-diverse humans. I don’t want to go too far into the details of the African genetics story, because 1. I’m convinced that it’s going to be my kind of must-see blockbuster, 2. I don’t want to spoil it based on only having seen a preview, and 3. the accelerating pace of discoveries in paleogenetics means it could be coming soon in all its high-def, wide-screen glory. Hopefully in plenty of time for me to know the story in my lifetime.
So knowing that I’m collapsing a lot of the complexity and richness we already have glimpsed, in broadest strokes, here is what we think we know about our second most numerous population of humans: not-very-diverse Sub-Saharan Africans. The founder population of today’s genetically not-very-diverse Sub-Saharan Africans shows signatures of a population crash, probably with a toll of 50% and probably from an effective breeding population of something like 50,000 people down to 25,000. So out of a founding population on the scale of a very small city 60,000-120,000 years ago, today’s 1.14 billion Sub-Saharan Africans have gone on to people every habitable space of that hugely varied continent. They are significantly more genetically diverse than their relatives on five other continents, but nowhere near as diverse as the neighbors in their midst.
What about the millions of humans around the world, and especially in the Western Hemisphere who carry both African and non-African ancestry? Well, that’s our fourth, intermediate group. If you are among the 140 or so million humans who are, for example, African-American, Brazilian, Caribbean, etc., look at your Ancestry, 23andMe, FamilyTree etc. results. The proportions in which you are genetically Sub-Saharan African vs. absolutely anything else are your ratio of not-very-diverse to very-not-diverse human ancestry. What if you simply have an 100% African parent and a 100% non-African parent? Then you don’t even need a genetic test. You are a 50-50 hybrid of not-very-diverse and very-not-diverse human lineages.
This is what diversity looks like
Finally, of course, we have that most unique group, our scant half-million genetically very-diverse relatives. People like the Khoi, the Mbuti, the San, those last thousand purely Hadza humans on the planet, seem to all be descended from a population whose crash was a little gentler. Our best guess today is that when they hit a bottleneck, they probably lost 25% of their prior diversity. So, if they only fell from an effective breeding population of 100,000 to 75,000, you can see why 60,000 to 120,000 years later, they are able to retain such a massively greater breadth of diversity within their genomes than all the other 7.7 billion of us combined.
What we thought we knew
In 2014, Richard Dawkins tweeted a picture of himself in a t-shirt that read “We Are All Africans.” Beside lending fertile basis for a meme (fitting, given the source), it is rooted in a central truth that genetics and paleoanthropology both have converged upon over the past 40 years: Homo sapiens is a “young” species with origins within the last few hundred thousand years on the continent of Africa.
In 1987, Allan Wilson’s group published a blockbuster result in Nature that gave rise to the popular conception of Mitochondrial Eve, the “mother of us all." Spun into a Nova documentary, Mitochondrial Eve reflected the discovery that all women could be shown to descend from a common ancestress around 200,000 years ago. This owed to the fact that all direct maternal lineages (recorded in our mitochondrial DNA (mtDNA) which passes from mother to daughter) can be traced back to a common genetic ancestor. Because most of the branches of the family tree of Eve wend their way back into Africa, the continent with the greatest mtDNA diversity, the team was able to conclude that our last common maternal ancestor was an ancient African.
From a paleontological finding of the earliest anatomically modern human in Ethiopia, the theory was further refined to have Eve living in East Africa. And from there, all of us, her descendants, were imagined to have spread to every corner of the globe. This, in turn, was followed and seemed substantiated by “Y-chromosomal Adam.” As with the mtDNA, the Y phylogeny’s deepest lineages lie within Africa.
Mitochondrial Eve was such a mainstay of our culture that it was featured in the series finale of Battlestar Galactica. But around this same time, in 2010, scientists discovered that Neanderthals, and a new human lineage in East Asia called “Denisovans”, contributed enduringly to the ancestry of all humans outside of Africa. The technological revolution of the Human Genome Project also yielded the complication that there seem to be wide differences in the magnitude and nature of the bottlenecks that had impacted different human populations.
Rather than a single punctuated event, where moderns suddenly became equipped to expand, replacing Neanderthals and overrunning other humans in a lone explosive pulse, the development of modern humanity was clearly more complex and multifocal. It occurred over hundreds of thousands of years of evolution within Africa, and later involved interbreeding with archaic hominins outside of Africa.
In short, the classic model was coming closest to fitting convincingly for non-Africans, while the complex situation within the mother continent still defied our simplifying impulses. Basically, our models for Africa remain more provisional.
But even as we home ever closer in on some clarity, we are still left with a host of unsolved mysteries. Why were the ancestors of non-Africans isolated for so long? Were they trapped on the other side of the Sahara from their African kin? If not there, where then? Arabia is another spot we can’t rule out. And why did they begin expanding explosively 60,000 years ago from a very small population? We still don’t know.
Below I answer some reader questions (Thanks for submitting these. Feel free to send along more.):
To what extent can we still describe Out of Africa as the basically correct account of humanity’s origins?
93-98.5% of the ancestry of humans outside of Sub-Saharan Africa (among those with no recent Sub-Saharan African ancestry, obviously) derives from a breeding population of 1,000 to 10,000, which expanded rapidly 60,000 years ago (reaching Australia and Europe around 45,000 and 50,000 years ago, accordingly).
So the simple “Out of Africa” narrative of a population crash and explosion across the world holds for North Africans, Eurasians, Oceanians, and Amerindians, some six and a half billion of us. But the origins of modern populations south of the Sahara are clearly more complex. Any bottleneck’s effects were much weaker within Africa, and multiple proto-modern populations seem to have been separating into distinct lineages as early as 200,000 years ago.
Additionally, Africa’s small populations of hunter-gatherers (that scant half-million people from above) are very different from agriculturalists (the other 1.14 billion Sub-Saharan Africans), the latter of whom are genetically closer to all the rest of humanity than to the hunter-gatherers near them on their same continent. We don’t have an exhaustive and comprehensive model to account for all these disparate facts. At least yet.
How does the newly defined Nesher Ramla Homo fit the picture?
We don’t know. The Nesher Ramla humans seem to have resembled European Neanderthals. If this surmise ends up holding, it is quite possible that the Neanderthal ancestry outside of Africa derives from this group.
Neanderthals occupied Europe, but they were also present as far east as western Mongolia. There are Neanderthal remains in Denisova cave, and a mixed Neanderthal-Denisovan girl was even discovered there. But Neanderthals were also present in West Asia, most famously in Shanidar Cave in Iraq. The Neanderthal ancestry in modern humans seems to be from a single group, and is present in all modern human populations from Europe to Australia. The most likely scenario is an admixture event in the Middle East then, as the nascent migration out of Africa ran into another human lineage and promptly absorbed it. If “Dragon Man” hadn’t randomly dropped at the same time, we’d probably be talking more about these possible Neanderthal ancestors of modern humans in Israel.
What is the status of the hypothesis that ancestors of Neanderthals and Denisovans developed outside Africa? (i.e. the scenario in Reich, page 69.). Any new evidence in the last few years?
We won’t get a good sense until we have ancient DNA or (more likely) proteins. Humans during the Pleistocene seem densest in Africa, then Southeast Asia, and finally the rest of Eurasia. So our most likely candidate is Africa. Paleoanthropologists still have more work to do.
Like the way the ~60kya 'out of Africa' populations spread and encountered/mixed with Neanderthals, Denisovans, etc., do you think there is a chance those Neanderthals/Denisovans, when they first spread out, intermixed with hominids leftover from even earlier out-of-Africa pulses?
It seems likely with Denisovans. Several research groups have found this result, that Denisovans have “deep” ancestry from other human populations. Probably, these would have been forms of Homo erectus that were present in East Asia when the Denisovans arrived more than 500,000 years ago. One of the diverse erectus populations could have been relatives of the Flores Hobbits, whose exact phylogenetic relationships are still somewhat unclear.
Do we know anything at all about what was so distinct about the group that came out of Africa 60,000 YA? Why them, why then, and what happened in Africa that empowered this group who left to sweep the rest of the globe?
We don’t know. As much of a non-answer as this is, it does seem like a recurrent pattern that human populations expand and replace local forerunners. The main thing we know about these proto-non-Africans (our very-not-diverse group from above) is that they went through a bottleneck where their effective breeding numbers crashed to 10% of the original population for many generations. Something like an original population of 10,000-100,000 dropping to 1000-10,000, with a value on the lower end looking likeliest in the modeling.
Of course, there are more advanced stone-tool technologies that emerge around this time. But, we know that modern humans are far older than these tool technologies, both due to the fossils as well as genetics, and we know they were already expanding, so it can’t just be the tools.
Are Khoi the only population without archaic introgressions?
No, they seem to have archaic introgression as well. Just from a “ghost” population from whom we don’t have an ancient genome.
Also do East Asians have Denisovan ancestry or not (conflicting accounts some say none, some say they have most)?
They certainly do have Denisovan ancestry. About 0.1% or so. Everyone east of the Indus has some, with most in the 0.1-0.3% interval, and a few, like Papuans, as high as 5%.
Homo Erectus in [Southeast Asia] also deeply structured?
Probably. These humans were in the area for over a million years, so the likelihood is that they diversified into lots of regional populations. Just like other apes in the region, such as gibbons and orangutans. We know Denisovans were deeply structured, so I see no reason to assume erectus was any different.
Is Out of Asia still possible in the development of sapiens?
Possible? Yes, I think so. But it would be a complex scenario, and seems to be improbable, though the possibility is greater than we’d have thought twenty years ago when modern human origins were defined by an exceedingly simple model. The old Out of Africa model, where humans expanded less than 100,000 years ago from a small tribe in East Africa, populated all of Africa, and then exited the continent to settle the rest of the world, excluded this scenario in totality. The newer models may have more room for the possibility, remote as it is.
How are Andaman Islanders closer to Swedes than to Africans?
The ancestral humans, who lost their fur two million years ago in Africa, probably had dark skin, so it seems like a good guess that the original population that moved out of Africa also had dark skin. Populations in the tropics maintained this darker skin, while those in northern latitudes developed lighter skin. The variation in physical appearance between non-Africans developed over the last 60,000 years. If you don’t have Sub-Saharan African ancestry, all of your ancestry (aside from your differential rates of Neanderthal/Denisovan ancestry) goes back to the same 1,000-10,000 humans that split off from other Africans. Just like every other person anywhere on the globe who doesn’t have any Sub-Saharan African ancestry.
People make the mistake of thinking Andaman Islanders are closer to Africans because of physical similarities. But this reflects a long-time tendency in our species to read too much into skin tone. As harmful as that simplistic bias has been over the centuries, it was more comprehensible when we lacked adequate scientific basis to understand our human family tree. In 2021, we know too much to let people make skin tone about something it’s not. Skin tone, like so many adaptations, is a faithful record of the conditions where your ancestors thrived. Period.
And actually, the genetic distance between a Papuan and a Nigerian is greater than between a Swede and a Nigerian. This is because Papuans have 5% Denisovan ancestry that Swedes (and Nigerians!) do not. Aside from that, their ancestral distance from Nigerians is the same. It all goes back to those 1,000-ish people 60,000 years ago.
Do we have adaptations from Neanderthals/Denisovans?
Yes. Hair, skin, and immune systems all seem to have been shaped by admixture with these human groups. There may even be circadian rhythm adaptations from Neanderthals in modern humans. Finally, there is also an association between certain types of autism and Neanderthal and Denisovan genes.
Were there migrations back to Africa during the Pleistocene?
Probably. But for that to fit the facts, the back-migrating population from Eurasia must have had no Neanderthal ancestry, which is at very low levels in Africa. It could be that the migration occurred before the Neanderthal admixture. Or, it could be that the immigrants were from an early group of non-Africans in the Middle East that never mixed with Neanderthals (this is basically the idea of “Basal Eurasians”).
Should we give up on ‘We Are All African’?
Replacement suggestion (and you can see why my t-shirt company went out of business): “We are all African and Neanderthal.”