Such a thing

“there’s no such thing as race” is a standard sentence in the United States and Europe. Conventional wisdom, and like so much conventional wisdom, false.

Of course there is.

First you need to define your terms. I would suggest that any population – a group whose members have mated within that group, almost entirely, for some time – and has experienced strong-enough natural selection to change significantly in some trait that we give a shit about can usefully be considered a race. Or a ‘goklu’, where goklu has exactly the same operational meaning as race, without having yet acquired any toxic associations. Low levels of inward gene flow allow selection to change the frequencies of alleles, so mating within the group is important. Usually this endogamy is a natural consequence of geography (not much gene flow across the Atlantic before Columbus) but sometimes it has been caused by social rules, as in the case of the Ashkenazi Jews or the Hindu castes.

low inward gene flow: in order for significant differences in the neutral genome to accumulate, there must have been < 1 immigrant per generation for tens of of thousands of years or more. That has happened sometimes, and not just with Neanderthals: sub-Saharan Africans and Eurasians were that separate until fairly recently, and have that kind of differences in their neutral genomes. For that matter, Bushmen and Bantu were genetically distinct for an even longer period. So it takes only a little gene flow to stop drift in its tracks.

Selection can be a lot stronger, and it takes more gene flow to scotch it. You could have effective selection for IQ among the Ashkenazi Jews even in the presence of as much as 0.5% inward gene flow per generation from the general European population. 2% would have been too much, though.

A long period of genetic isolation does not automatically generate differences in any particular trait: but it does show that there has been an extended opportunity for selection to operate effectively and generate population differences.

So when we see differences, how old are they? and how can we tell? Plausible selection pressures could generates one-std trait differences in as little as a thousand years, and in some cases, like the Ashkenazim, it likely has. In other cases it may have operated over tens of thousands of years, even as much as quarter of a million years (Bushmen/Pygmies versus other humans).

If the trait in question is characteristic of a geographically extended population, you might suspect that selection had operated over a long time. But since we now know that there have been many population expansions and replacements, you might be wrong. Ancient DNA may be a better guide.

So sometimes the explanation for the differences between two populations may go back deep into the Ice Age, but it might also have happened since the birth of agriculture, or even since the fall of Rome.

Suppose you have a one-std difference in some trait between two populations? What can we say about the genetic architecture? Well, sometime it boils down to the presence or absence of a single allele. Other times it is caused by a shift in the frequencies of a number of alleles that each have a small effect on the trait.

African-Americans average about 1-std lower in white count. That’s all due to the Duffy allele. All else equal, northern Europeans are a couple of centimeters taller than southern Europeans: that is caused by frequency differences in hundreds of alleles affecting height, a shift that on the whole has increased the frequency of plus variants.

So what to say to someone that asks about the ‘race gene’? First, you tell her that she’s an idiot. The complex of shovel-shaped incisors, thick hair, small breasts, more eccrine sweat glands, and a different shape to the hangy-down part of the ear, fixed in northeast Asia, is indeed caused by a single allele, an EDAR variant that is essentially nonexistent in Europe or Africa. On the other hand, Pygmy height, or the lack of it, is influenced by a number of alleles.

But the genetic architecture isn’t all that important: it’s the differences that matter. Pygmies are really short – that’s what matters.

Along those lines, Lewontin and other bullshit artists have tried to argue that genetic statistics are such that human groups can’t really be different. Most genetic variation in humans is within-group, rather than between-group: so fucking what? the same is true for dogs: am I supposed to think that pit bulls and Chihuahuas and border collies are ‘really the same’?

Having more plus variants in the alleles that affect a particular quantitative trait doesn’t show up in these genetic statistics (like Fst) at all. Neither would a big frequency difference in a single allele that had a big effect, like EDAR.

People are mostly about as different as they seem to be. There are exceptions, cases where an environmental insult makes a fair amount of difference. This is particularly the case with height, where nutritional status can easily create a 1-std difference. But height is influenced by genetics, too, and the shortest people (the Pygmies) are short for genetic reasons, not because they’re starving.

What about the magic immunity of the brain to natural selection? That’s nonsense, of course. We know, for sure, that different goklus have different distributions of personality traits – because they act significantly differently with 24 hours of birth. All the psychometric results indicate that goklus vary in intelligence too [perhaps 3 stds from highest to lowest] probably largely because of differences in the frequency of many alleles with small effects.

Posted in Uncategorized | 148 Comments


If you had a reasonably cheap particle accelerator, one that could produce a high flux of few-GeV neutrinos, you could make yourself some real money.

Posted in Uncategorized | 50 Comments

Bugs Versus Drift

For a while now we have seen occasional articles about how people outside of sub-Saharan Africa may have more genetic load, generated by drift over a population history in which Eurasians generally had a smaller effective population size. This is related to those recent papers with similar but stronger conclusions about Eurasian archaic humans like Neanderthals and Denisovans.

As I mentioned, I kind of doubt that that Neanderthals and Denisovans were all that screwed up – partly because a more complete theory indicates that salvage mutations get easier as you drift away from the optimum and ameliorate the effects of low population size to a surprising degree, but even more because the Neanderthals obviously weren’t all that screwed up, didn’t have their fitness depressed by tens of percent, because they stood off anatomically modern humans for something like 70,000 years after first encounter. Results count. The mechanism makes a lot more sense for Flores hobbits because their population was much smaller, but even more because they actually were screwed up: it shows in their skeletons.

So, just how screwed up do Eurasians look, compared to Africans? I mean, if we’re going to be busy explaining a phenomenon, shouldn’t we bother to make a cursory check to see whether it even exists? I know that probably sounds radical…

Of course there’s no bloody sign of any such thing. Sub-Saharan Africans have shorter lifespans and lower IQs than most Eurasian populations. East Asia has lower genetic diversity than Europeans – so has had lower effective population size over the past few tens of thousands of years – yet those populations have higher average IQs and longer lifespans than Europeans.

On the other hand, it is also the case that strong selection for any particular trait tends to mess up other traits – logical, considering trade-offs. So strong selection for resistance to falciparum malaria has made lots of deleterious variants common in the tropical and subtropical parts of the Old World. We know many that are strongly, obviously bad for you – even lethal – but there are surely many others with milder (but still negative) effects. We know that selection for resistance to sleeping sickness has selected for APOL1 variants that greatly increase the risk of kidney failure, so that African-Americans develop end-stage renal disease (ESRD) about 3.5 times more often than people of European ancestry. Those APOL1 variants also exacerbate kidney disease caused by sickle cell, and there’s an AIDs-related kidney-wrecking syndrome (HIVAN: HIV-associated collapsing glomerulopathy) (while AIDs is of course much more common in blacks) that seems to require those African APOL1 variants – HIVAN is up to 50 times more common in blacks than whites. Those social constructs can sure seem real when they’re shoveling the dirt over your face.

Posted in Uncategorized | 91 Comments

The 1%

We don’t see people today with Neanderthal Y chromosomes or mtDNA. I keep hearing people argue that this means that mating between Neanderthal males and AMH females must have produced sterile males, or that matings between AMH men and Neanderthal women were all sterile, or whatever.

That is not necessarily the case. A slight disadvantage is all that would be required to totally eliminate Neanderthal Y-chromosomes or mtDNA.

Imagine that a Neanderthal Y-chromosome reduces the bearer’s fitness by 1%, and that the original frequency of Neanderthal Y chromosomes (after admixture) was 2%.

It’s been something like 1500 generations. The expected frequency is 5.67 x 10-9. In real life it would probably have fluctuated to zero, and of course stayed there.

Understand and remember.

Posted in Uncategorized | 41 Comments

The Long Count

They’ve managed to sequence a bit of autosomal DNA from the Atapuerca skeletons, about 430,000 years old, confirming that they are on the Neanderthal branch.

Among other things, this supports the slow mutation rate, one compatible with what we see in modern family trios, but also with the fossil record.

This means that the Pygmies, and probably the Bushmen also, split off from the rest of the human race about 300,000 years ago. Call them Paleoafricans.

They are anatomically modern: they have chins, etc. Behaviorally modern? There have been only a few attempts to measure their intelligence: what has been done indicates that they have very low IQs. They definitely talk, tell stories, sing songs: does that imply that they could, given the right environment, have developed the Antikythera mechanism or a clipper ship?

This means that language is older than some had thought, a good deal older. It also means that people with language are quite capable of going a quarter of a million years without generating much technological advance – without developing the ability to push aside archaic humans, for example. Of course, people with Williams syndrome have language, and you can’t send them into the kitchen and rely on them to bring back a fork. Is the sophistication of Bushman language – this means the concepts they can and do convey, not the complexity of the grammar – comparable with that of other populations? I don’t know. As far as I can see, one of the major goals of modern anthropology is to make sure that nobody knows. Or that they know things that aren’t so.

The minimal definition of behavioral modernity – that set of traits that exists in all of humanity, including those that are most divergent, and that are probably ancestral in anatomically modern humans – may not include much technological creativity.

Next: since we now know that generic Neanderthal and Denisovan alleles don’t fit too well with anatomically modern humans, and AMH alleles didn’t fit too well with Altai Neanderthals, it seem likely that you see the beginning of such functional divergence between Paleoafricans and everyone else. I know of one example of a European haplotype that’s a heart disease risk on a mostly-African genetic background, but not on a European background, but most such incompatibilities are probably very mild, hard to detect. It would probably take thousands of generations for a Pygmy population to lose a significant fraction of its Bantu introgression. You might be able to detect this on a few alleles, but for the most part it just hasn’t been long enough.

Culture gets forgotten, inventions get lost: any populations with a sufficiently low innovation rate probably does not advance at all, culturally. They could respond to natural selection, change in a way that increased their innovation rate… So people could continue to make Acheulean handaxes for a million years: they had to change before there could be further technological progress. Neanderthals had more sophisticated technology, but that technology changed very, very slowly compared to, say , that of humans in the upper Paleolithic. Biology keeps culture on a leash, and you can get to the end of the leash.

Some have suggested that the key to technological development is higher population: that produces more intellects past a high threshold, sure. I don’t think that’s the main factor. Eskimos have a pretty advanced technology, but there were never very many of them. On the other hand, they have the highest IQ of any existing hunter-gatherer population: that’s got to help. Populations must have gone up the Eemian, the previous interglacial period, but nothing much got invented back then. It would seem that agriculture would have been possible in the Eemian, but as far as we know it didn’t happen. Except for Valusia of course. With AMH going back at least 300,000 years, we have to start thinking about even earlier interglacial peiods, like Mindel-Riss (424-374 k years ago)

Posted in Uncategorized | 40 Comments

Henry Harpending

Henry Harpending has died.

He suffered a stroke 3 weeks ago. Within a few days, he also had a MRSA infection in his lungs. The docs eventually cleared that, but his lungs never recovered. He died this afternoon of Adult Respiratory Distress Syndrome.

Posted in Uncategorized | 132 Comments

Two Kinds of Indians

There’s a new paper out on Amerindian genetic history in Science,mostly looking at 92 ancient mitochondrial genomes (500 to 8600 years old). Because mtDNA has a high mutation rate, it’s useful for seeing rapid changes in population size (Bayesian skyline analysis). It looks as if the Amerindian wave started about 16k years ago, first down the Pacific coast, since this was before the ice-free passage across the Rockies opened up.

There is that weird Andaman-like signature; which they say is compatible with a second migration. They also say that it might have come later, but looking at the distribution, that’s impossible. There’s no way to come in later and end up spread out at the 2% level all across the Amazon. More later.

They see clear evidence of a big population expansion after the move into the Americas – of course that makes sense.

People don’t seem to have moved around too much after the initial settlement. No big expansions like you see in the Old World.

Now for the weirdness. None of the ancient DNA samples have any descendants or close relatives today. Zip. “no ancient haplotype shares a common ancestor with a modern haplotype more recently than ~9,000 years ago. Their samples were mostly taken from large population centers along the western coast of South America: apparently those people went extinct after Columbus. The model that best fit their data was the following:

“Model C also assumes one panmictic population of constant size (2,000 females), and the same growth between 500–600 generations ago as in Model A. Following the massive increase in Ne we assumed a split into two demes (0 and 1) of equal size (50,000 females each with no migration between demes) around 360 generations ago (~9,000 years ago or the last time we observe an ancestor common to modern and ancient lineages in hg A2; black triangle in Fig. 3B), and a constant population size for both (‘geographically’) isolated demes. Ancient lineages were exclusively sampled from deme 1 by converting radiocarbon or archeological dates into number of generations. To account for an extinction of ancient lineages as a consequence of the contact with European colonizers, we modeled a drastic population decline from 50,000 to 1 female (virtually extinct) for deme 1 between 20 generations ago (~1500 AD) and to the present day (time 0). In contrast, we assumed an exponential growth for deme 0 to 100,000 females over the last 20 generations to account for the present-day mitochondrial diversity. ”

They posit two distinct kind of Indians: those that left fossil DNA, and those that left descendants. Good Indians, and bad Indians.

Now it makes sense that lowland types suffered more than people in the Andean or Mexican highlands: they faced African diseases like falciparum malaria and yellow fever as well as the huge array of Eurasian diseases. And the Spanish had their own problems with altitude in the Andean highlands.

But this is nuts! or at any rate amazing. It suggests that in large areas, Amerindians went extinct, and were later replaced by different Amerindians from other places (plus Spaniards and Portuguese), not too closely related. That’s not incredibly far from what happened in the Caribbean, maybe: we know the local Indians disappeared rapidly there (although even there, they are a significant maternal component in existing populations). And there are indications that at least some of the hunter-gatherer tribes of the Amazon might be new to the neighborhood.. And maybe densely populated areas were hit hardest by the new infectious diseases, with a few of the Amerindian equivalent of hillbillies surviving…

And maybe there is undersampling of existing populations. But it sure sounds weird.

“Our ancient samples were principally derived from large population centers along the western coast of South America, which experienced high extinction rates following European colonization. Historic accounts have reported that the population decline was more rapid and intense in the Gulf of Mexico and the Pacific coast of Peru than in other areas such as the Mesoamerican plateau or the Andean highlands (7).

Posted in Uncategorized | 33 Comments