Tag: Genetics

Quantitative genomics, adaptation, and cognitive phenotypes

Posted on by Razib Khan

The human brain utilizes about ~20% of the calories you take in per day. It’s a large and metabolically expensive organ. Because of this fact there are lots of evolutionary models which focus on the brain. In Richard Wrangham suggests that our need for calories to feed our brain is one reason we started to use fire to pre-digest our food. In Geoffrey Miller seems to suggest that all the things our big complex brain does allows for a signaling of mutational load. And in Robin Dunbar suggests that it’s social complexity which is driving our encephalization.

These are all theories. Interesting hypotheses and models. But how do we test them? A new preprint on bioRxiv is useful because it shows how cutting-edge methods from evolutionary genomics can be used to explore questions relating to cognitive neuroscience and pyschopathology, Polygenic selection underlies evolution of human brain structure and behavioral traits:

…Leveraging publicly available data of unprecedented sample size, we studied twenty-five traits (i.e., ten neuropsychiatric disorders, three personality traits, total intracranial volume, seven subcortical brain structure volume traits, and four complex traits without neuropsychiatric associations) for evidence of several different signatures of selection over a range of evolutionary time scales. Consistent with the largely polygenic architecture of neuropsychiatric traits, we found no enrichment of trait-associated single-nucleotide polymorphisms (SNPs) in regions of the genome that underwent classical selective sweeps (i.e., events which would have driven selected alleles to near fixation). However, we discovered that SNPs associated with some, but not all, behaviors and brain structure volumes are enriched in genomic regions under selection since divergence from Neanderthals ~600,000 years ago, and show further evidence for signatures of ancient and recent polygenic adaptation. Individual subcortical brain structure volumes demonstrate genome-wide evidence in support of a mosaic theory of brain evolution while total intracranial volume and height appear to share evolutionary constraints consistent with concerted evolution…our results suggest that alleles associated with neuropsychiatric, behavioral, and brain volume phenotypes have experienced both ancient and recent polygenic adaptation in human evolution, acting through neurodevelopmental and immune-mediated pathways.

The preprint takes a kitchen-sink approach, throwing a lot of methods of selection at the phenotype of interest. Also, there is always the issue of cryptic population structure generating false positive associations, but they try to address it in the preprint. I am somewhat confused by this passage though:

Paleobiological evidence indicates that the size of the human skull has expanded massively over the last 200,000 years, likely mirroring increases in brain size.

From what I know human cranial sizes leveled off in growth ~200,000 years ago, peaked ~30,000 years ago, and have declined ever since then. That being said, they find signatures of selection around genes associated with ‘intracranial volume.’

There are loads of results using different methods in the paper, but I was curious note that schizophrenia had hits for ancient and recent adaptation. A friend who is a psychologist pointed out to me that when you look within families “unaffected” siblings of schizophrenics often exhibit deviation from the norm in various ways too; so even if they are not impacted by the disease, they are somewhere along a spectrum of ‘wild type’ to schizophrenic. In any case in this paper they found recent selection for alleles ‘protective’ of schizophrenia.

There are lots of theories one could spin out of that singular result. But I’ll just leave you with the fact that when you have a quantitative trait with lots of heritable variation it seems unlikely it’s been subject to a long period of unidirecitional selection. Various forms of balancing selection seem to be at work here, and we’re only in the early stages of understanding what’s going on. Genuine comprehension will require:

– attention to population genetic theory
– large genomic data sets from a wide array of populations
– novel methods developed by population genomicists
– and funcitonal insights which neuroscientists can bring to the table

South Asian gene flow into Burmese and Malays?

Posted on by Razib Khan


I happen to have a data set merged from the 1000 Genomes and Estonian Biocentre which has Malays, Burmans, and other assorted Southeast Asians, East Asians, and South Asians. In light of recent posts I thought I would throw out something in relation to this data set (you can download the data here). Above you can see the populations in the data. You see Bangladeshis consistently are shifted toward Southeast Asians in comparison to other South Asians. But both Burmans and Malays exhibit some shift toward South Asians.

I ran ADMIXTURE at K = 4. Click the image for the larger file which shows the populations, but I will tell you what’s going on.

The yellow to green represent a north-south axis in East Asia. The Han sample is mostly yellow, but there is a green component in varying degrees. This almost certainly represents heterogeneity in the Han sample of north to south Chinese. The green component is nearly ~100% in some individuals from indigenous tribes in Borneo, and balanced with the yellow among peninsular Malays. It is more at a higher frequency in Cambodia than in Vietnam or Burma, indicating the older roots of Khmers and their relative insulation from later migrations of Sino-Tibetan and Tai peoples.

The red South Asian component is found in many Southeast Asians, but curious in the Burmans and Malays there is a lot of variation within the population. That indicates admixture over time that has not homogenized throughout the population.

I ran Treemix with 5 migration edges and French rooted (1000 SNP blocks out of 225,000 SNPs) and they all looked like this. Commentary I will leave to readers….

Genetics books for the masses!

Posted on by Razib Khan

Since I’ve become professionally immersed in genetics I haven’t read many books on the topics. I read papers. And I do genetics. But back in the day I did enjoy a good book. The standard recommendation would be to read Matt Ridley’s . It’s a bit dated now (it was published around when the Human Genome Project being completed), but I’d still recommend it.

But when in the mid-2000s I dabbled a little bit in the world of worm (C. elegans) genetics I read Andrew Brown’s . It’s pretty far from my current concerns and fixations, with more of a focus on developmental processes, but it is pretty cool to read about the race to “map” every cell in C. elegans.

The second book I’d recommend readers of this blog is the late Will Provine’s . Modern population genomics is a massive edifice built atop the foundations of the early 20th century fusion of Mendelism and the biometrical heirs of Darwin. Provine outlines how primitive genetics eventually seeded the birth of the Neo-Darwinian Synthesis.

Why do percentage estimates of “ancestry” vary so much?

Posted on by Razib Khan

When looking at the results in , , and my “East Asian” percentage is:

– 19%
– 13%
– 6%

What’s going on here? In science we often make a distinction between precision and accuracy. Precision is how much your results vary when you re-run an experiment or measurement. Basically, can you reproduce your result? Accuracy refers to how close your measurement is to the true value. A measurement can be quite precise, but consistently off. Similarly, a measurement may be imprecise, but it bounces around the true value…so it is reasonably accurate if you get enough measurements just cancel out the errors (which are random).

The values above are precise. That is, if you got re-tested on a different chip, the results aren’t going to be much different. The tests are using as input variation on 100,000 to 1 million markers, so a small proportion will give different calls than in the earlier test. But that’s not going to change the end result in most instances, even though these methods often have a stochastic element.

But what about accuracy? I am not sure that old chestnuts about accuracy apply in this case, because the percentages that these services provide are summaries and distillations of the underlying variation. The model of precision and accuracy that I learned would be more applicable to the DNA SNP array which returns calls on the variants; that is, how close are the calls of the variant to the true value (last I checked these are arrays are around 99.5% accurate in terms of matching the true state).

What you see when these services pop out a percentage for a given ancestry is the outcome of a series of conscious choices that designers of these tests made keeping in mind what they wanted to get out of these tests. At a high level here’s what’s going on:

  1. You have a model of human population history and dynamics with various parameters
  2. You have data that that varies that you put into that model
  3. You have results which come back with values which are the best fit of that data to the model you specificed

Basically you are asking the computational framework a question, and it is returning its best answer to the question posed. To ask whether the answer is accurate or not is almost not even wrong. The frameworks vary because they are constructed by humans with difference preferences and goals.

Almost, but not totally wrong. You can for example simulate populations whose histories you know, and then test the models on the data you generated. Since you already know the “truth” about the simulated data’s population structure and history, you can see how well your framework can infer what you already know from the patterns of variation in the generated data.

Going back to my results, why do my East Asian percentages vary so much? The short answer is that one of the major variables in the model alluded to above is the nature of the reference population set and the labels you give them.

Looking at Bengalis, the ethnic group I’m from, it is clear that in comparison to other South Asian populations they are East Asian shifted. That is, it seems clear I do have some East Asian ancestry. But how much?

The “simple” answer is to model my ancestry is a mix of two populations, an Indian one and an East Asian one, and then see what the values are for my ancestry across the two components. But here is where semantics becomes important: what is Indian and East Asian? Remember, these are just labels we give to groups of people who share genetic affinities. The labels aren’t “real”, the reality is in the raw read of the sequence. But humans are not capable of really getting anything from millions of raw SNPs assigned to individuals. We have to summarize and re-digest the data.

The simplest explanation for what’s going on here is that the different companies have different populations put into the boxes which are “Indian/South Asian” and “East Asian.” If you are using fundamentally different measuring sticks, then there are going to be problems with doing apples to apples comparisons.

My personal experience is that 23andMe tends to give very high percentages of South Asian ancestry for all South Asians. Because “South Asian” is a very diverse category when tests come back that someone is 95-99% South Asian…it’s not really telling you much. In contrast, some of the other services may be using a small subset of South Asians, who they define as “more typical”, and so giving lower percentages to people from Pakistan and Bengal, who have admixture from neighboring regions to the west and east respectively.*

Something similar can occur with East Asian ancestry. If the “donor” ancestral groups are South Asian and East Asian for me, then the proportions of each is going to vary by how close the donor groups selected by the company is to the true ancestral group. If, for example,  chose a more Northeastern Asian population than , then my East Asian population would vary between the two services because I know my East Asian ancestry is more Southeast Asian.

The moral of the story is that the values you obtain are conditional on the choices you make, and those choices emerge from the process of reducing and distilling the raw genetic variation into a manner which is human interpretable. If the companies decided to use the same model, the would come out with the same results.

* I helped develop an earlier version of MyOrigins, and so can attest to this firsthand.

But evolution converges!

Posted on by Razib Khan

Stephen Jay Gould became famous in part for his book . By examining the strange creatures in the Burgess Shale formation Gould makes the case that evolution is a highly contingent process, and that if you reran the experiment of life what we’d see might be very different from what we have now.

But the scientist whose study of the formation that inspired Gould’s interpretation, Simon Conway Morris, had very different views. Though it can sometimes be churlish, his rebuttal can be found in . Simony Conway Morris does not believe that contingency is nearly as powerful a force as Gould would have you believe. And his viewpoints are influential. Richard Dawkins leaned on him to make the case for convergence in evolution in .

This crossed my mind when reading Carl Zimmer’s new column, When Dinosaurs Ruled the Earth, Mammals Took to the Skies:

Today, placental mammals like flying squirrels and marsupials like sugar gliders travel through the air from tree to tree. But Volaticotherium belonged to a different lineage and independently evolved the ability to glide.

They were not the only mammals to do so, it turns out. Dr. Luo and his colleagues have now discovered at least two other species of gliding mammals from China, which they described in the journal Nature.

…

Dr. Meng said that the growing number of fossil gliders showed that many different kinds of mammals followed the same evolutionary path. “They did their own experiments,” he said.

This ultimately comes down to physics. There are only so many ways you can make an organize that flies or glides. Mammals come to the table with a general body plan, and that can be modified only so many different ways.

This is not a foolproof point of datum in favor of convergence as opposed to contingency. Frankly these are often vague verbal arguments which are hard to refute or confirm. And even molecular evolutionary analyses come to different conclusions. It may be that we are asking the wrong question. But, it does suggest that evolution may work in a much narrower range of parameters as time progresses because of the winnowing power of selection.

Jon Snow + Daenerys Targaryen far creepier genetically than you know

Posted on by Razib Khan
Screenshot 2016-06-14 22.09.51
Credit: poly-m (deviantART)

If you have been following you have been noticing that there is a brewing romance between Jon Snow, King in the North, and Daenerys Targaryen, the aspiring claimant to her father’s Iron Throne.

Of course there is a twist to all of this: unbenknownst to either, Jon Snow’s biological father is Daenerys’ dead brother, Rhaegar. This means that Daenerys is Jon Snow’s aunt.

Long-time followers of the world of  are aware that incest between near relations is neither unknown nor shocking. But there is a non-trivial detail which it is important to note. Jon and Daenerys are far more closely related than typical aunts and nephews.

The reason is simple, Daenerys and her brother were the products of two generations of sibling incest. Incest results in inbreeding, and inbreeding as you know results in loss of genetic diversity. By Daenerys’ generation the coefficient of relationship between herself and her brothers was much higher than normal.

To be concrete, the coefficient of relationship of full-siblings is 0.50. That of half-siblings 0.25. Identical twins? Obviously 1.0. Another way to think about this is how much of the genome do any two pairs of individuals share in terms of long tracts of inheritance from recent ancestors. On the whole siblings share about half of their genomes in such a fashion. After two generations of inbreeding Daenerys and Rhaegar have a coefficient of relationship of 0.727 (using Wright’s method). They’re not identical twins, obviously, but their genetic relationship is far closer than full-siblings!

Don’t let the mother of dragons ride you Jon!

Dividing  this in half gives 0.36 as the coefficient of relationship between Jon and Daenerys, as opposed to 0.50 for full-siblings and 0.25 for a conventional aunt-nephew. Jon and Daenerys have almost the same genetic relationship as 3/4 siblings; two individuals who share a common parent, like half-siblings, but whose unshared parents are first order relatives (full-siblings or parent-child).

Not Jaime & Cersei creepy, but still creepy.

Addendum: Though Daenerys is quite inbred, Jon is not at all. One generation of outbreeding can eliminate all inbreeding.

When the ancestors were cyclops

Posted on by Razib Khan


The Greeks are important because Western civilization began with Greece. And therefore modern civilization. I don’t think the Greeks were “Western” truly; my own preference is to state that the West as we understand it is really just Latin Christendom, which emerged in the late first millennium A.D. in any coherent fashion. Yet without Classical Greece and its accomplishments the West wouldn’t make any sense.

But here I have to stipulate Classical, because Greeks existed before the Classical period. That is, a people who spoke a language that was recognizably Greek and worshipped gods recognizable to the Greeks of the Classical period. But these Greeks were not proto-Western in any way. These were the Mycenaeans, a Bronze Age civilization which flourished in the Aegean in the centuries before the cataclysms outlined in

The issue with the Mycenaean civilization is that its final expiration in the 11th century ushered in a centuries long Dark Age. During this period the population of Greece seems to have declined, and society reverted to a more simple structure. By the time the Greeks emerged from this Dark Age much had changed. For example, they no longer used Linear B writing. Presumably this technique was passed down along lineages of scribes, whose services were no longer needed, because the grand warlords of the Bronze Age were no longer there to patronize them and make use of their skills. In its stead the Greeks modified the alphabet of the Phoenicians.

To be succinct the Greeks had to learn civilization all over again. The barbarian interlude had broken continuous cultural memory between the Mycenaeans and the Greeks of the developing polises of the Classical period. The fortifications of the Mycenaeans were assumed by their Classical descendants to be the work of a lost race which had the aid of monstrous cyclops.

Of course not all memories were forgotten. Epic poems such as retained the memory of the past through the centuries. The list of kings who sailed to Troy actually reflected the distribution of power in Bronze Age Greece, while boar’s tusk helmets mentioned by Homer were typical of the period. To be sure, much of the detail in Homer seems more reflective of a simpler society of petty warlords, so the nuggets of memory are encased in later lore accrued over the centuries.

When antiquarians and archaeologists began to take a look at the Bronze Age Aegean the assumption by many was that the Mycenaeans were not Greek, but extensions of the earlier Minoan civilization. The whole intellectual history here is outlined in Michael Wood’s 1980s documentary . But suffice it to say that many were shocked when Michael Ventris deciphered Linear B, and found that it was clearly Greek!

The surprise here was partly due to the fact that though Mycenaean cultural remains indicated a different civilization from that of the Minoans, its motifs are clearly inherited from the earlier group. Mycenaeans seemed in many ways to be Minoans in chariots. And the presumption has long been that the Minoans themselves were not an Indo-European population. In fact, the island of Crete had developed early on and become part of the orbit of civilized states from the northern Levant down to Egypt, including Cyprus. Therefore some scholars hypothesized an Egyptian connection.

In any case, the Mycenaeans were Greek. And Homer then most certainly must have transmitted traditions which went back to the Bronze Age.

At this point we can now speak to demographics with some data, as Nature has come out with a paper using ancient DNA from Mycenaeans, Minoans, as well as Bronze Age Anatolians, Genetic origins of the Minoans and Mycenaeans:

The origins of the Bronze Age Minoan and Mycenaean cultures have puzzled archaeologists for more than a century. We have assembled genome-wide data from 19 ancient individuals, including Minoans from Crete, Mycenaeans from mainland Greece, and their eastern neighbours from southwestern Anatolia. Here we show that Minoans and Mycenaeans were genetically similar, having at least three-quarters of their ancestry from the first Neolithic farmers of western Anatolia and the Aegean1, 2, and most of the remainder from ancient populations related to those of the Caucasus3 and Iran4, 5. However, the Mycenaeans differed from Minoans in deriving additional ancestry from an ultimate source related to the hunter–gatherers of eastern Europe and Siberia6, 7, 8, introduced via a proximal source related to the inhabitants of either the Eurasian steppe1,6, 9 or Armenia4, 9. Modern Greeks resemble the Mycenaeans, but with some additional dilution of the Early Neolithic ancestry. Our results support the idea of continuity but not isolation in the history of populations of the Aegean, before and after the time of its earliest civilizations.

About 85% of the ancestry of the Minoan samples could be modeled as being derived from Anatolian farmers, the ancestors of the “Early European Farmers” (EEF) that introduced agriculture to most of the continent, and whose heritage is most clear in modern populations among Sardinians. For the three Mycenaean samples the value is closer to 80% (though perhaps high 70s is more accurate).

Now the question though is what’s the balance? For the Minoans the residual is a component which seems to derive from “Eastern Farmer” populations. Additionally the authors note that the Y chromosomes in four out of five individuals in their Mycenaean-Minoan-Anatolians are haplogroup J associated with these eastern groups, rather than the ubiquitous G2 of the earlier farmer populations. The authors suggest that in the 4th millennium B.C. there was a demographic event where this ancestral component swept west, and served as the common Mycenaean-Minoan (and Anatolian) substrate.

But the Mycenaean samples (one of which was elite, two of which were not) also have a third component: affinities with steppe populations. One model which presents itself is that there was a pulse out of the Balkans, and this was part of the dynamic described in Massive migration from the steppe was a source for Indo-European languages in Europe. But another model, which they could not reject, is that the steppe affinity came from the east, perhaps from a proto-Armenian population. Additionally, they did not find much steppe ancestry in the Anatolian samples at all.

My own preference is for a migration through the Balkans. It seems relatively straightforward. As for why the Anatolian samples did not have the steppe ancestry, the authors provide the reasonable supposition that Indo-European in Anatolia branched off first, and the demographic signal was diluted over successor generations. Perhaps. But another aspect of Anatolia is that it seems the Hittites, the Nesa, where never a numerous population in comparison to the Hatti amongst whom they lived. Perhaps a good model for their rise and takeover may be that of the post-Roman West and the Franks in Gaul.

Then the question becomes how does a less numerous people impose their language on a more numerous one? This happens. See the Hungarians for an example. In fact the paper which covered the other end of the Mediterranean, The population genomics of archaeological transition in west Iberia: Investigation of ancient substructure using imputation and haplotype-based methods, suggests that language shift can occur in unpredictable ways. On the one hand Basques seem to have mostly Indo-European Y chromosomes, but their whole genome ancestry indicates less exogenous input than their neighbors. Speaking of which, we know by the Classical period large regions of western Spain were dominated by Celtic speaking peoples, but  the genetic imprint of the Indo-Europeans is still very modest in the Iberian peninsula.

I think what we’re seeing here is the difference between Indo-European agro-pastoralists arriving to a landscape of relatively simple societies with more primal institutions, and those who migrated into regions where local population densities are higher and social complexity is also greater. This higher social complexity means that external elites can takeover a system, as opposed to an almost animal competition for resources as seems to have occurred in Northern Europe.

Finally, at the end of the supplements there is an analysis of the physical features of the Minoans and Mycenaeans. There’s not much that’s surprising. The Minoans and Mycenaeans were a dark haired and dark eyed folk. Why should this surprise us at all? We actually have self representations of them! That’s what they look like. If anything they were darker than modern Greeks (small sample size means power to draw conclusions is not high). Why?

Two reasons that come to mind: natural selection, and the fact that modern Greeks seem to be shifted to continental Europeans to their north, likely due to migration. My number one contender here are the Scalveni Slavic tribes which pushed into much of Greece in the second half of the 6th century A.D. (though a minority of Greek samples I’ve seen don’t exhibit much skew toward Slavs at all).

In the future with more samples and more genomes we’ll know more. But I think this work emphasizes that when it comes to Europe most of the demographic patterns we see around us date to the Bronze Age or earlier.

The future will be genetically engineered

Posted on by Razib Khan


If the film had come out a few years later I believe there would have been mention of CRISPR. Sometimes science leads to technology, and other times technology aids in science. On occasion the two are one in the same.

The plot I made above shows that in the first five years of the second decade of the 20th century CRISPR went from being an obscure aspect of bacterial genetics to ubiquitous. Friends who had been utilizing “advanced” genetic engineering methods such as TALENS and zinc fingers switched overnight to a CRISPR/Cas9 framework.

As I’ve said before the 2010s are the decade when “reading” the genome becomes normal. We really don’t know what the CRISPR/Cas9 technology is capable of. It’s early years yet. With that, First Human Embryos Edited in U.S.. Technically they’re single celled zygotes. The science itself is not astounding. Rather, it is that the human rubicon has been passed in the United States. As indicated in the article there has been some jealousy about what the Chinese have been able to do because of a different cultural and regulatory framework.

There are those calling for a moratorium on this work (on humans). I’m not in favor or opposed. Rather, my question is simple: if CRISPR/Cas9 makes genetic engineering cheap, easy, and effective, how exactly are we going to enforce a world-wide moratorium? A Butlerian Jihad?

Note: I know that people are freaking about humans + genetic engineering. But most geneticists I know are more excited about the prospects of non-human work, since human clinical trials are going to be way in the future. Over 20 years since Dolly it’s notable to me that no human has been cloned from adult somatic cells yet.

Indian genetic history: before the storm

Posted on by Razib Khan

Over at Brown Pundits I’ve mentioned the continuing simmer of controversy over a recent piece, How genetics is settling the Aryan migration debate. This has prompted responses in the Indian media from a Hindu nationalist perspective. One of these notes that the author of the piece above cites me, and then goes on to observe I was fired from The New York Times a few years ago due to accusations of racism (also, there is the implication that I’m just a blogger and we should trust researchers with credibility like Gyaneshwer Chaubey; well, perhaps he should know that Gyaneshwer Chaubey considers me “unbiased” according to an email exchange which I had with him last week [we all have biases, so I think he’s wrong in a literal sense]).

I was a little surprised that a right-wing magazine would lend legitimacy to the slanders of social justice warriors, but this is the world we live in. Those who believe that everything written about me in the media, I invite you to submit your name and background to me. I have contacts in the media and can get things written if I so choose. Watch me write something which is mostly fact, but can easily be misinterpreted by those who Google you, and watch how much you value the objective “truth-telling” power of the press all of a sudden.

There’s a reason so many of us detest vast swaths of the media, though to be fair we the public give people who don’t make much money a great deal of power to engage in propaganda. Should we be surprised they sensationalize and misrepresent with no guilt or shame? I have seen most of those who snipe at me in the comments disappear once I tell them that I know what their real identity is. Most humans are cowards. I have put some evidence into the public record to suggest that I’m not.

Perhaps more strange for me is that the above piece was passed around favorably by , who I was on friendly terms with (we had dinner & drinks in Brooklyn a few years back). I asked him about the slander in the piece and he unfollowed me on Twitter (a friend of Hindu nationalist bent asked Sanjeev on Facebook about the articles’ attack on me, but the comment was deleted). It shows how strongly people feel about these issues.

I’m in a weird position because I’m brown and have a deep interest in Indian history. But that interest in Indian history isn’t because I’m brown, I’m pretty interested in all the major zones of the Old World Oikoumene. Aside from some jocular R1a1a chauvinism I don’t have much investment personally (I just told said Hindu nationalist friend who turns out to be R2 to clean my latrine; joking of course, though I’m sure he resents that I’m descended on the direct paternal line from the All-Father & Lord of the Steppes and he is not!).

In the aughts I accepted the model outlined in 2006’s The Genetic Heritage of the Earliest Settlers Persists Both in Indian Tribal and Caste Populations. But to be frank it always struck me as a little confusing because the tentative autosomal data we had suggested that many South Asians were closer to West Eurasians than deep divergences dating to the Last Glacial Maximum would suggest. Since I’ve written something like 5 million words in 15 years, I actually can check if I’m remembering correctly. So here’s a post from 2008 where I express reservations of the idea of long term deep heritage of Indians separate from other West Eurasians. The reason I was so impressed by 2009’s Reconstructing Indian Population History is that it resolved the paradox of South Asian genetic relatedness.

To recap, Reich et al. proposed that modern Indians (South Asians) could be modeled as a two way mixture between two distinct populations with separate evolutionary genetic histories, Ancestral North Indians and Ancestral South Indians (ANI and ASI). How distinct? ANI were basically another West Eurasian population, while ASI was likely nested in the clade with Eastern Non-Africans. Additionally, there was a NW-to-SE and caste admixture cline. In other words, the higher you were on the caste ladder the more ANI you had, and the closer your ancestors were from the north and west, and more ANI you had. The difference between Y and mtDNA, male and female, could be explained by sex-biased migration.

But there were still aspects of the paper which I had reservations about. After all, it was a model.

  • Models are imperfect fits onto reality. The idea of mass migration seemed ridiculous to me at the time, because even by the time of the Classical Greeks it was noted that India was reputedly the most populous land in the world (to their knowledge). But ancient DNA has convinced me of the reality of mass migrations.
  • I wasn’t sure about the nature of the closest modern populations to the ANI. The researchers themselves (in particular, Nick Patterson) told me that the relatedness of ANI to Europeans was very close (on the order of intra-European differences). But modern Indians do not look to be descended from a population that is half Northern European physically. Again, ancient DNA has shown that there was lots of population turnover, and it turns out that Europeans and ANI were likely both compounds and mixed daughter populations of common ancestors (also, typical European physical appearance seems to have emerged in situ over the past 5,000 years).
  • The two way admixture modeled seemed too simple. I had run some data and it struck me that North Indian populations like Jats had something different than South Indian groups like Pulayars. In 2013 Priya Moorjani’s paper pretty much confirmed that it was more than a two way admixture along the ANI-ASI cline.

This March BMC Evolution Biology published Silva et al’s A genetic chronology for the Indian Subcontinent points to heavily sex-biased dispersals. It has made a huge splash in India, arguably triggering the write up in The Hindu. But for me it was a bit ho-hum. If you read my 2008 post it is pretty clear that I suspected the most general of the findings in this paper at least 10 years back. It is nice to get confirmation of what you suspect, but I’m more interested to be surprised by something novel.

Nevertheless A genetic chronology for the Indian Subcontinent points to heavily sex-biased dispersals has come in for lots of repeated attack in the right-wing Indian press. This is unfair, because it is a rather good paper. I suspect that it wasn’t published in a higher ranked journal because most scientists don’t consider the history of India to be that important, and they didn’t really apply new methods, as opposed to bringing a bunch of data and methods together (in contrast, the 2009 Reich et al. paper was one of the first publications which showed how to utilize “ghost populations” in explicit phylogenetic models with relevance to human demographic history).

As it happens I will be writing up my thoughts in detail in an article for a major Indian publication (similar circulation numbers as The Hindu). This has been in talks for over six months, but I’ve been busy. But a month or so ago I thought it was time that I put something into print for the Indian audience, because I felt there was some misrepresentation going on (i.e., the Aryan invasion theory has not been been refuted by genetics, but this is what many Indians assert).

For any years people have told me there are certain topics that shouldn’t be talked about. I have offended people greatly. There are many things people do not want to know. I have come to the conclusion this is not an entirely indefensible viewpoint (though if you accept this viewpoint, I think acceptance of authoritarianism is inevitable, so I hope people will toe the line when the new order arrives; knowing their personalities I think they will conform fine). But my nature is such that I continue to have nothing but contempt for the duplicitous and craven manner in which people go about these sorts of private conversations. I assume that as someone with the name “Razib Khan” I will be attacked vociferously by Hindu nationalists, who will no doubt make recourse to the Left-wing hit pieces against me to undermine my credibility. The fact that these groups are fellow travelers should tell us something, though I will leave that as an exercise for the reader.

I will write my piece that reflects the science as I believe it is, without much consideration of the attacks. That is rather easy for me to do in part because I live in the United States, where denigrating the deeply held views and self-esteem of Hindu nationalists is not sensitive or politically protected (unlike say, Muslims). And Hindu nationalists are less likely to kill me by orders of magnitude than Muslim radicals, and they have far less purchase in this nation then the latter (though you may be interested to know that very conservative Muslims follow me on ; they’re actually more open-minded than many SJWs to be entirely honest).

Let me go over some general points that I see coming up over and over on the relationship between Indian (pre)history and genetics in the critiques .

One of the major critiques has to do with the nature of R1a-Z93 and its subclades. Basically this Y chromosomal haplogroup, the greatest that has ever been known, exhibits a strong signature of very rapid expansion over the past 4,000 years or so. It is divided from Z282. While Z93 is found in South Asia, Central Asia, and Siberia, Z282 is European, with its dominant subclade the one associated with Eastern Europeans. Both of these clades of R1a have gone through massive expansion. In the Altai region R1a is 40% of the heritage of peoples who are now predominantly East Eurasian today. But they are Z93. Additionally, ancient DNA from the Pontic Steppe dated ~4,000 years ago from Srubna remains is Z93, as are Scythian remains from the Iron Age.

Much of the argument comes down to dating, and citing papers that give deep coalescence numbers between difference branches of R1a1a. Hindu nationalists and their fellow travelers point to recent papers which give dates >10,000 years ago, and so place the origin of Z93 plausibly in the Pleistocene. The problem is that Y chromosomal coalescence dating is something of a mug’s game. Often they use microsatellite data whose mutational rates are highly uncertain. In contrast, using SNP data, which has a slower mutation rate but requires a lot more data, you get TRMCA (common ancestry) between Z93 and Z282 around ~5,800 years ago. But coalescence estimates often have wide confidence intervals of thousands of years. And even with these intervals, the assumptions you make (e.g., mutation rate) strongly influence your midpoint estimate.

The Y chromosomal data is powerful, but its interpretation is still buttressed upon other assumptions. The really big picture framework is the nature of ancient genome-wide variation across Eurasia. Lazaridis et al. 2016 condition us to a prior where much of Eurasia was subject to massive population-wide genetic changes since the Holocene. Therefore, I am much less surprised if there was massive genetic change in India relatively recently. The methods in Priya Moorjani’s paper and in other publications make it obvious that mixture was extensive in South Asia between very distinct groups until about ~2,000 years ago. In fact, Moorjani et al. using patterns of variation across the genome to come at a number of two to four thousand years ago as the period of massive admixture.

Though we don’t have relevant ancient DNA from India proper to answer any questions yet, we do have ancient DNA from across much of Europe, Central Asia, and the Near East. What they show is that Indian populations share ancestry from both Neolithic Iranians and peoples of the Pontic steppe, who flourished ~5 to ~10,000 years ago. To some extent the latter population is a daughter population of the former…which makes things complicated. Conversely, no West Eurasian population seems to harbor ancient signals of ASI ancestry.

One scientist who holds to the position that most South Asian ancestry dates to the Pleistocene argued to me that we don’t know if ancient Indian samples from the northwest won’t share even more ancestry than the Iranian Neolithic and Pontic steppe samples. In other words, ANI was part of some genetic continuum that extended to the west and north. This is possible, but I do not find it plausible.

The reasons are threefold. First, it doesn’t seem that continuous isolation-by-distance works across huge and rugged regions of Central Eurasia. Rather, there are demographic revolutions, and then relative stasis as the new social-cultural environment crystallizes. This inference I’m making from ancient DNA and extrapolating. This may be wrong, but I would bet I’m not off base here.

Second, it strikes me as implausible that there was literally apartheid between ASI and ANI populations for the whole Holocene right up until ~4,000 years before the present. That is, if Northwest India was involved in reciprocal gene flow with the rest of Eurasia over thousands of years I expect there should have been some distinctive South Asian ASI-like ancestry in the ancient DNA we have. We do not see it.

Third, one of the populations with strong affinities to some Indian populations are those of the Pontic steppe. But we know that this group itself is a compound of admixture that arose 5,000-6,000 years ago. Because of the complexity of the likely population model of ANI this is not definitive, but it seems strange to imagine that ANI could have predated one of the populations with which it was in genetic continuum as part of a quasi-panmictic deme.

Finally, many of the critiques involve evaluation of the scientific literature in this field. Unfortunately this is hard to do from the outside. Citing papers from the aughts, for example, is not wrong, but evolutionary human population genomics is such a fast moving field that even papers published a few years ago are often out of date.

Many are citing a 2012 paper by a respected group which argues for the dominant model of the aughts (marginal population movement into South Asia). One of their arguments, that Central Asian migrant should have East Asian ancestry, is a red herring since it is well known that this dates to the last ~2,000 years or so (we know more now with ancient DNA). But the second point that is more persuasive in the paper is that when they look at local ancestry of ANI vs. ASI in modern Indians, the ANI haplotypes are more diverse than West Eurasians, indicating that they are  not descendants but rather antecedents (usually the direction of ancestry is from more diverse to less due to subsampling).

There are two points that I have make here. First, local ancestry analysis is difficult, so I would not be surprised if they integrated ASI regions into ANI and so elevated the diversity in that way (though they think they’ve taken care of it in the paper). Second, if the ANI are a compound of several West Eurasian groups then we expect them to be more diverse than their parents. In other words, the paper is refuting a model which is almost certainly incorrect, but the alternative hypothesis is not necessarily the true hypothesis (which is a more complex demographic model than many were testing in 2012).

But there are many things we do not know still. Many free variables which we haven’t nailed down. Here are some major points:

  • Y chromosomal lineages have a correlation with ethno-linguistic groups, but the correlation is imperfect. R1b and R1a seems correlated with Indo-European groups, but both these are found in high proportions in groups which are putatively mostly “pre-Indo-European” in origin (e.g., Basques, Sardinians, and South Indian tribals and non-Brahmin Dravidian speaking groups). Also, haplogroups like I1 in Europe expand with Indo-Europeans locally, suggesting there was lots of heterogeneity in Indo-Europeans as they expanded. In other words, Indo-European expansion in relation to powerful paternal lineages did not always correlate with ethno-linguistic change.
  • There are probably at minimum two Holocene intrusions from the northwest into South Asia, but this is a floor. The models that are constructed always lack power to detect more complexity. E.g., it is not impossible that there were several migrations of Indo-Europeans into South Asia which we can not distinguish genetically over a period of a few thousand years.
  • If one looks over all of South Asia it may be that ASI ancestry in totality is >50% of the total genome ancestry. I don’t have a good guess of the numbers. If this is correct, perhaps most South Asian ancestors 10,000 years ago were living in South Asia (though the fertility rate are such in Pakistan that ANI ancestry is increasing right now in relative rates).
  • But, this presupposes that ASI were present in South Asia in totality 10,000 years ago, rather than being migrants themselves. If ancient DNA confirms that ANI were long present in Northwest India, I hold then it is entirely likely that ASI was intrusive to South Asia! The BMC Evolutionary Biology Paper does a lot of interpretation of deep structure in haplogroup M in South Asia. I’m moderately skeptical of this. Europe may not be a good model for South Asia, but there we see lots of Pleistocene turnover.

So where does this leave us? Ancient DNA will answer a lot of questions. Pretty much all scientists I’ve talked to agree on this. My predictions, some of which I’ve made before:

  1. The first period of admixture is old, and dates to the founding of Mehrgarh as an agricultural settlement. The dominant ANI component dates to this period and mixture event, all across South Asia. The presence in South India is due to expansion of these farming populations.
  2. A second admixture event occurred with the arrival of steppe people. Those who argue for the Aryan invasion model posit 1500 BCE as the date. But these people probably were expanding in some form before this date.
  3. We still don’t know who the antecedents for the Indo-Aryans were. Probably they were a compound of different steppe groups, and also other populations which were mixed in (by analogy, in Europe it is obvious now that there was some mixture with the local European farmers and hunter-gatherers as Europeans expanded their frontier westward; the same probably applies for Indo-Aryans are the BMAC).

The fad for dietary adaptations is not going away

Posted on by Razib Khan


Food is a big deal for humans. Without it we die. Unlike some animals (here’s looking at you pandas) we’re omnivorous. We eat fruit, nuts, greens, meat, fish, and even fungus. Some of us even eat things which give off signals of being dangerous or unpalatable, whether it be hot sauce or lutefisk.

This ability to eat a wide variety of items is a human talent. Those who have put their cats on vegetarian diets know this. After a million or so years of being hunters and gatherers with a presumably varied diet for thousands and thousands of years most humans at any given time ate some form of grain based gruel. Though I am sympathetic to the argument that in terms of quality of life this was a detriment to median human well being, agriculture allowed our species to extract orders of magnitude more calories from a unit of land, though there were exceptions, such as in marine environments (more on this later).

Ergo, some scholars, most prominently Peter Bellwood, have argued that farming did not spread through cultural diffusion. Rather, farmers simply reproduced at much higher rates because of the efficiency of their lifestyle in comparison to that of hunter-gatherers. The latest research, using ancient DNA, broadly confirms this hypothesis. More precisely, it seems that cultural revolutions in the Holocene have shaped most of the genetic variation we see around us.

But genetic variation is not just a matter of genealogy. That is, the pattern of relationships, ancestor to descendent, and the extent of admixtures across lineages. Selection is also another parameter in evolutionary genetics. This can even have genome-wide impacts. It seems quite possible that current levels of Neanderthal ancestry are lower than might otherwise have been the case due to selection against functional variants derived from Neanderthals, which are less fitness against a modern human genetic background.

The importance of selection has long been known and explored. Sickle-cell anemia only exists because of balancing selection. Ancient DNA has revealed that many of the salient traits we associate with a given population, e.g., lactose tolerance or blue eyes, have undergone massive changes in population wide frequency over the last 10,000 years. Some of this is due to population replacement or admixture. But some of it is due to selection after the demographic events. To give a concrete example, the frequency of variants associated with blue eyes in modern Europeans dropped rapidly with the expansion of farmers from the Near East ~10,000 years ago, but has gradually increased over time until it is the modal allele in much of Northern Europe. Lactase persistence in contrast is not an ancient characteristic which has had its ups and downs, but something new that evolved due to the cultural shock of the adoption of dairy consumption by humans as adults. The region around lactase is one of the strongest signals of natural selection in the European genome, and ancient DNA confirms that the ubiquity of the lactase persistent allele is a very recent phenomenon.

But obviously lactase is not going to be the only target of selection in the human genome. Not only can humans eat many different things, but we change our portfolio of proportions rather quickly. In a the economic historian Gregory Clark observed that English peasants ate very differently before and after the Black Death. As any ecologist knows populations are resource constrained when they are near the carrying capacity, and England during the High Medieval period there was massive population growth due to gains in productivity (e.g., the moldboard plough) as well as intensification of farming and utilization of all the marginal land.

After the Black Death (which came in waves repeatedly) there was a massive population decline across much of Europe. Because institutions and practices were optimized toward maintaining a much higher population, European peasants lived a much better lifestyle after the population crash because the pie was being cut into far fewer pieces. In other words, centuries of life on the margins just scraping by did not mean that English peasants couldn’t live large when the times allowed for it. We were somewhat pre-adapted.

Our ability to eat a variety of items, and the constant varying of the proportions and kind of elements which go into our diet, mean that sciences like nutrition are very difficult. And, it also means that attempts to construct simple stories of adaptation and functional patterns from regions of the genome implicated in diet often fail. But with better analytic technologies (whole genome sequencing, large sample sizes) and some elbow grease some scientists are starting to get a better understanding.

A group of at Cornell has been taking a closer look at the FADS genes over the past few years (as well as others at CTEG). These are three nearby genes, FADS1, FADS2, and FADS3 (they probably underwent duplication). These genes are involved in the metabolization of fatty acids, and dietary regime turns out to have a major impact on variation around these loci.

The most recent paper out of the Cornell group, Dietary adaptation of FADS genes in Europe varied across time and geography:

Fatty acid desaturase (FADS) genes encode rate-limiting enzymes for the biosynthesis of omega-6 and omega-3 long-chain polyunsaturated fatty acids (LCPUFAs). This biosynthesis is essential for individuals subsisting on LCPUFA-poor diets (for example, plant-based). Positive selection on FADS genes has been reported in multiple populations, but its cause and pattern in Europeans remain unknown. Here we demonstrate, using ancient and modern DNA, that positive selection acted on the same FADS variants both before and after the advent of farming in Europe, but on opposite (that is, alternative) alleles. Recent selection in farmers also varied geographically, with the strongest signal in southern Europe. These varying selection patterns concur with anthropological evidence of varying diets, and with the association of farming-adaptive alleles with higher FADS1 expression and thus enhanced LCPUFA biosynthesis. Genome-wide association studies reveal that farming-adaptive alleles not only increase LCPUFAs, but also affect other lipid levels and protect against several inflammatory diseases.

The paper itself can be difficult to follow because they’re juggling many things in the air. First, they’re not just looking at variants (e.g., SNPs, indels, etc.), but also the haplotypes that the variants are embedded in. That is, the sequence of markers which define an association of variants which indicate descent from common genealogical ancestors. Because recombination can break apart associations one has to engage with care in historical reconstruction of the arc of selection due to a causal variant embedded in different haplotypes.

But the great thing about this paper is that in the case of Europe they can access ancient DNA. So they perform inferences utilizing whole genomes from many extant human populations, but also inspect change in allele frequency trajectories over time because of the density of the temporal transect. The figure to the left shows variants in both an empirical and modeling framework, and how they change in frequency over time.

In short, variants associated with higher LCPUFA synthesis actually decreased over time in Pleistocene Europe. This is similar to the dynamic you see in the Greenland Inuit. With the arrival of farmers the dynamic changes. Some of this is due to admixture/replacement, but some of it can not be accounted for admixture and replacement. In other words, there was selection for the variants which synthesize more LCPUFA.

This is not just limited to Europe. The authors refer to other publications which show that the frequency of alleles associated with LCPUFA production are high in places like South Asia, notable for a culture of preference for plant-based diets, as well as enforced by the reality that animal protein was in very short supply. In Europe they can look at ancient DNA because we have it, but the lesson here is probably general: alternative allelic variants are being whipsawed in frequency by protean shifts in human cultural modes of production.

In  Lawrence Keeley observed that after the arrival of agriculture in Northern Europe in a broad zone to the northwest of the continent, facing the Atlantic and North Sea, farming halted rather abruptly for centuries. Keeley then recounts evidence of organized conflict in between two populations across a “no man’s land.”

But why didn’t the farmers just roll over the old populations as they had elsewhere? Probably because they couldn’t. It is well known that marine regions can often support very high densities of humans engaged in a gathering lifestyle. Though not farmers, these peoples are often also not nomadic, and occupy areas as high density. The tribes of the Pacific Northwest, dependent upon salmon fisheries, are classic examples. Even today much of the Northern European maritime fringe relies on the sea. High density means they had enough numbers to resist the human wave of advance of farmers. At least for a time.

Just as cultural forms wane and wax, so do some of the underlying genetic variants. If you dig into the guts of this paper you see much of the variation dates to the out of Africa period. There were no great sweeps which expunged all variation (at least in general). Rather, just as our omnivorous tastes are protean and changeable, so the genetic variation changes over time and space in a difficult to reduce manner. The flux of lifestyle change is probably usually faster than biological evolution can respond, so variation reducing optimization can never complete its work.

The modern age of the study of natural selection in the human genome began around when A Map of Recent Positive Selection In the Human Genome was published. And it continues with methods like SDS, which indicate that selection operates to this day. Not a great surprise, but solidifying our intuitions. In the supplements to the above paper the authors indicate that the focal alleles that they are interrogating exhibit coefficients of selection around ~0.5% or so. This is rather appreciable. The fact that fixation has not occurred indicates in part that selection has reversed or halted, as they noted. But another aspect is that there are correlated responses; the FADS genes are implicated in many things, as the authors note in relation to inflammatory diseases. But I’m not sure that the selection effects of these are really large in any case. I bet there are more important things going on that we haven’t discovered or understood.

Obviously genome-wide analyses are going to continue for the foreseeable future. Ten years ago my late friend Mike McKweon predicted that at some point genomics was going to have be complemented by detailed follow up through bench-work. I’m not sure if we’re there yet, but there are only so many populations you can sequence, and only to a particular coverage to obtain any more information. Some selection sweeps will be simple stories with simple insights. But I suspect many more like FADS will be more complex, with the threads of the broader explanatory tapestry assembled publications by publication over time.

Citation: Ye, K., Gao, F., Wang, D., Bar-Yosef, O. & Keinan, A. Dietary adaptation of FADS genes in Europe varied across time and geography. Nat. Ecol. Evol. 1, 0167 (2017).