A guide to the science and pseudoscience of A Troublesome Inheritance, part III: Has natural selection produced significant differences between races?

This is the third in a series of guest posts in which Chris Smith will examine the evolutionary claims made in Nicholas Wade’s book A Troublesome Inheritance. You can read part I here, and part II here. Chris is an Associate Professor of Evolutionary Ecology at Willamette University. He uses population genetic approaches to understand coevolution of plants and insects, and he teaches the interdisciplinary course “Race, Racism, and Human Genetics” with Emily Drew.

A Troublesome Inheritance was published in 2014 by Penguin Books. Cover image via Google Books.

This spring former New York Times science writer, Nicholas Wade, released his latest book on human evolution, A Troublesome Inheritance: Genes, Race, and Human History. In it, Wade argues that genetic studies completed in the eleven years since the Human Genome Project was completed reveal real and important differences between human races. Unsurprisingly, the book’s release has been met with a sharply divided critical reception.Whereas the book has been widely embraced by those on the political right, and by the white identity movement, it has been panned by anthropologists, evolutionary biologists, and population geneticists. For the last two weeks at Nothing in Biology Makes Sense, I’ve been looking in depth at the literature that Wade uses to support his ideas. Last week I considered Wade’s argument that natural selection acting on the MAO-A gene – a neurotransmitter implicated in aggression and impulsivity – has led to behavioral differences between races. This week I will consider Wade’s larger claim that natural selection has produced numerous differences between races.

Throughout the book Wade continually repeats the mantra that natural selection on humans has been “recent, copious, and regional.” It would be hard to find an evolutionary biologist that would disagree with these rather vague pronouncements. Indeed, there are a multitude of studies showing that natural selection has acted on humans, and there is persuasive evidence that selection has caused evolutionary changes in human populations as we have adapted to diverse environments over the course of the last several thousand years (see, for example, Yi et al., 2010).

However, scratching the surface reveals that when he says that natural selection has been “recent, copious, and regional,” what Wade actually means is that natural selection has been “radical, complete, and racial.” By Wade’s account, natural selection has dramatically reshaped the human genome, producing major differences between races. This much more dramatic interpretation is entirely unsupported by the literature, however. In truth, Wade vastly overstates the portion of the human genome that shows evidence for natural selection, and where there has been recent natural selection acting on humans, its effect has primarily been to create genetic differences between members of the same race, and similarities between people of different races.

Just how “copious” are we talkin’ here?

In support of his view that natural selection has made major revisions to the human genome, Wade repeatedly cites the statistic that, “No less than 14% of the genome” has been changed by recent natural selection. This claim, however, is such a staggering misinterpretation of the evidence that I am almost inclined to think Wade must not have read more than the abstracts of most of the papers he cites. The ‘14%’ figure comes from a recent review by Jonathon Akey, who compared 21 separate studies that searched the entire human genome for evidence of natural selection (Akey, 2009). What Akey found was there was little agreement among the various studies about which parts of the genome have experienced natural selection. And although the number ‘14%’ does appear in this review, that number refers to sum total of all regions of the genome identified as showing evidence of selection in any one study.

So, for example, if one study one identified selection acting on gene A, but not gene B, and another study found selection acting on gene B, but not gene A, the ‘14% of the genome’ that Wade references would refer to genes A and B together. But Akey’s point in citing this dramatic number is to call attention to the wide disagreement among studies. Akey writes, “In total, 5110 distinct regions were identified in one or more study. These regions encompass 409 Mb of sequence (14% of the genome) and contain… 23% of all genes… Strikingly, only 722 regions (14.1% [of all genes]) were identified in two or more studies, 271 regions (5.3% [of all genes]) were identified in three or more studies, and 129 regions (2.5% [of all genes]) were identified in four or more studies.”

So, in fact, the Akey study that Wade cites in support of his “copious” natural selection headline, actually shows that only a tiny fraction of the genome has been consistently identified as being under selection. And although there is potentially a ‘glass half-full’ interpretation of Akey’s findings – there are, after all, hundreds of regions identified by multiple studies as showing evidence of selection, and some of the disagreement between studies may be due to different methodological approaches used by the different investigators – the story that Akey’s review tells is actually one of widespread disagreement among the various studies. And that’s not just my interpretation; Akey wrote as much in the original paper saying, “There is no escaping the general conclusion that the overlap among studies is under-whelming.”

In addition, regardless of which figure we use – 23%, 14%, or 2.5% – the proportion of the genome identified as showing evidence of natural selection is almost certainly vastly greater than the portion of the genome that was actually the focus of selection. The reason this is true is because of phenomenon known as ‘genetic hitchhiking.’ Briefly, genes are not free-floating in the cell like so many plastic nibs in a beanbag, but instead are organized onto chromosomes – long linear stretches of DNA. Some genes may be located very nearby, “linked” to one another on the chromosome, and these linked genes tend to be passed on together from parent to offspring. As a result, selection acting on a particular locus (that is, natural selection favoring a particular “allele” or genetic variant of a gene) often affects other nearby, linked genes as well. Depending on how strongly natural selection has acted on a gene, and how recently this evolutionary change occurred, very large sections of the chromosome, containing many linked genes, may also be affected. The reason this is important is that often a particular gene may statistically ‘look like’ it has been the target of selection, even though its evolution was actually shaped by genetic hitchhiking instead of natural selection[1].

A ‘selective sweep.’

Under natural selection, a new beneficial mutation will rise in frequency (prevalence) in a population. A schematic shows polymorphisms along a chromosome, including the selected allele, before and after selection. Ancestral alleles are shown in grey and derived (non-ancestral) alleles are shown in blue. As a new positively-selected allele (red) rises to high frequency, nearby linked alleles on the chromosome ‘hitchhike’ along with it to high frequency, creating a ‘selective sweep.’

Regional does not mean racial

So much for the ‘copious’ part of Wade’s argument. What about his claim that natural selection has been “regional” Here, again, Wade’s failure to understand some important details lead him to conclusions that are wholly unsupported by the literature he cites. Although it is certainly true that humans have adapted to different environments as we dispersed across the globe, the effect of these adaptations have been to create differences between populations within races (say, between Finns and Italians), not to create differences between races (say, between Africans and Asians).

For example[2], an exciting recent study compared genetic variation between people living in Beijing, Tibet, and the Netherlands, and found very strong evidence that natural selection caused rapid evolutionary change among Tibetans in a gene called EPAS-1, which is believed to play a role in the development of red blood cells (Yi et al., 2010). According to the authors of this study, the differences in EPAS-1 they identified are “likely to confer a functionally relevant adaptation to the hypoxic environment of high altitude.” For our purposes, however, the important point here is that the effect of natural selection here was to create differences between Tibetans and people living in Beijing, both of whom would be categorized as Asian under any conventional racial classification.

Indeed, one of the reasons we can be confident that the differences in EPAS-1 are due to natural selection (as opposed to other evolutionary processes, such as genetic drift) is that with respect to this gene people living in Beijing (Asians) are much more similar to residents of the Netherlands (Europeans) than they are to people living in Tibet. This, despite the fact that the distance from Beijing to Tibet is about the same as from Chicago to Los Angeles, while the Dutch live almost a hemisphere away.

On the flip side, natural selection has also acted to created similarities between people of different races—both genetic similarities and similarities in phenotype, an individual’s observable characteristics. For example, differences in skin color are very likely to be the result of selection due to differences in the intensity of ultraviolet variation (UVA and UVB light) across the Earth’s surface. Across the globe differences in human skin pigmentation are very strongly associated with the intensity of UV light, and where there are mismatches (i.e., where people with very dark skin live in regions that receive little UV radiation, or people with very light skin that live in regions with intense sunlight), these are almost always due to recent migration (for example, English colonists of Australia).

Variation in human skin pigmentation within “races”, as defined by Wade.

Variation in human skin pigmentation within “races”, as defined by Wade. Each bar shows the total range in human skin color from most lightly pigmented (the highest reflectance, or the top of each bar) to most darkly pigmented (the lowest reflectance, or the bottom of each bar). There is substantial variation in skin color within races, with Europeans embracing all of the variation in skin color found within East Asians, and most of the variation found within Africans. This figure is based on raw data described in Jablonski & Chaplin (2000), who measured light reflectance at 685 nm. I assigned each population measured by Jablonski and Chaplin to one of the three ‘major races’ described by Wade. For these purposes I omitted data from Australians and Papuans, and from Native Americans, which Wade describes as ‘minor races’. If these were included, they would be grouped with Asians, the effect of which would be to substantially increase the phenotypic variation within Asians, as some native Australians have skin that is more darkly pigmented than any Africans’.

So, despite the fact that skin color is the feature most commonly used to identify someone’s race, there is both substantial variation in skin color within the geographic groups that Wade identifies as races, and in many cases natural selection has independently produced similar skin tones in people of different races (Jablonski & Chaplin, 2010). For example, people living in the sub-continent of India and in central and south America appear to have independently evolved darkly pigmented skin as they adapted to regions where there are high levels of ultraviolet radiation (Jablonski & Chaplin, 2010). Similarly, selection for improved production of vitamin D in higher latitudes, where insolation is lower, has caused European and east Asian populations to independently evolve lighter skin through different genetic mechanisms (Norton et al., 2007).

Another example of natural selection acting on humans, one that will be familiar to many graduates of introductory biology courses, is resistance to malaria. Throughout much of human history, malaria was an endemic human blood parasite, which is often reported (perhaps erroneously) to have caused half of all human deaths. The plasmodium parasite that causes malaria is transmitted to its human host by a mosquito bite, eventually invading red blood cells where it grows and divides. Classical work in human genetics has shown that a variant form of human hemoglobin (the HbS allele of the β-globin gene) confers partial resistance to malaria; when a blood cell is infected by the malaria plasmodium the hemoglobin molecule deforms, causing the blood cell carrying it to become deformed and elongated. These ‘sickle-shaped’ cells are then broken down by the spleen, eliminating the malaria parasite in the process (Roth et al., 1978; Wellems, Hayton, & Fairhurst, 2009). Within regions where malaria is common natural selection has acted to maintain this variant form of hemoglobin at high frequency in human populations (Wellems et al., 2009).


The distribution of malaria parasites (area shaded in hatched lines) versus the relative frequency of the sickle cell allele in human populations (red shading; dark red having the highest frequency, light red having the lowest). Image from “Connection Between DNA and Phenotype” in Principles of Biology, an OpenStax online textbook by Robert Bear and David Rintoul.

However, malarial resistance also comes at a cost; the resulting destruction of red blood cells can result in a condition known as ‘sickle-cell anemia’, in which the sickle-shaped cells block the flow of blood through capillaries, and limit blood flow to organs and limbs. Typically, sickle-cell anemia is thought of as a classically ‘racialized’ disease, because of its prevalence in Africans and people of recent African descent (Tapper, 1999). However, as with skin color, the effect of selection for resistance to malaria has actually been to create differences between people within races, and similarities between people of different races. In fact, sickle-cell disease and the associated hemoglobin variants are restricted to the equatorial regions within Africa, and are almost completely absent from southern African and the cape region (Piel et al., 2013). On the other hand, sickle-cell anemia is common in the Persian Gulf and eastern India (Weatherall & Clegg, 2001). Finally, the variant form of hemoglobin is far from the only mechanism of resistance to malaria (Wellems et al., 2009), and other genetic variants associated with malaria resistance are also shared between people of different races (Weatherall & Clegg, 2001).

A last example which will be familiar to many readers of Nothing In Biology, but which may be less well-known to those outside of the evolutionary biology sphere, is the ability to digest milk as an adult. Amongst mammals, humans are unusual in that (some of us) continue to drink milk after weaning (Itan, Jones, Ingram, Swallow, & Thomas, 2010). Because most mammals stop drinking milk relatively early in life, the body stops producing lactase – the enzyme needed to digest milk sugar, or lactose. As a result, most ‘normal’ mammals are lactose intolerant as adults. However, among human populations that have developed the cultural practice of keeping dairy cattle, some people continue to produce lactase as adults (Tishkoff et al., 2007). Remarkably, this evolutionary innovation has occurred repeatedly (Itan et al., 2010; Tishkoff et al., 2007), with natural selection favoring the independent evolution of ‘lactase persistence’ in different populations. Examining the global distribution of lactase persistence genes, we see that here, again, the effect of natural selection has been to create similarities between people of different “races,” and differences between people within races.

Distribution of malaria.

Geographic variation in the frequency of lactase persistence, the ability to digest milk sugar as an adult, after weaning. Dots represent collection locations. Colors and color key show the frequencies of the lactase persistence phenotype estimated by surface interpolation. Image from Itan et al. (2010)

So, in summary what we see is that although humans have clearly been shaped by natural selection, the extent of selection is far less than what Wade contends. Indeed, other population genetic forces such as genetic drift and population expansion may have played a larger role in our evolution. Furthermore, when we examine the effects of natural selection overall its impact has been to cause different populations to adapt to local environmental conditions. That is, because the environment varies much more within continents than between continents (think, for example, of the difference between the climate in Northern Canada and Baja California, or between Thailand and Siberia), adaptation through natural selection has produced differences between people living in different environments, and similarities between people in similar environments, even if the live on different continents. By and large natural selection has NOT produced major differences between “races.”

I say “Tomato”, you say “Race”

Undoubtedly to some readers the distinction here must seem like semantics. If natural selection has in fact produced differences between people – albeit differences between local populations, and not between broad continental races – doesn’t this whole argument come down to what you mean by “race”? And if all we’re fighting about is how best to carve up humanity (Should we recognize three races? Or 14? Or more?), can’t we just do what Ella Fitzgerald recommended and call the whole thing off?

In fact, however, the difference is not just an aesthetic preference for one classification scheme over another. Alternative conceptions of race have major consequences both for how we understand human diversity from a scientific perspective, and for very practical matters such as how doctors assess a patient’s risk factors for developing certain diseases and even how these illnesses are treated. Next week, in the final installment of this series, I will consider the question of how genetic and phenotypic variation is distributed within humans, and how alternative definitions of race have major, real-world consequences.


Akey J.M. (2009). Constructing genomic maps of positive selection in humans: Where do we go from here?, Genome Research, 19 (5) 711-722. DOI: 10.1101/gr.086652.108

Itan Y., Catherine JE Ingram, Dallas M Swallow & Mark G Thomas (2010). A worldwide correlation of lactase persistence phenotype and genotypes, BMC Evolutionary Biology, 10 (1) 36. DOI: 10.1186/1471-2148-10-36

Jablonski N. (2000). The evolution of human skin coloration, Journal of Human Evolution, 39 (1) 57-106. DOI: 10.1006/jhev.2000.0403

Jablonski N.G. (2010). Human skin pigmentation as an adaptation to UV radiation, Proceedings of the National Academy of Sciences, 107 (Supplement_2) 8962-8968. DOI: 10.1073/pnas.0914628107

Norton H.L., E. Parra, P. McKeigue, X. Mao, K. Cheng, V. A. Canfield, D. G. Bradley, B. McEvoy & M. D. Shriver (2006). Genetic evidence for the convergent evolution of light skin in Europeans and East Asians, Molecular Biology and Evolution, 24 (3) 710-722. DOI: 10.1093/molbev/msl203

Piel F.B., Rosalind E Howes, Oscar A Nyangiri, Peter W Gething, Mewahyu Dewi, William H Temperley, Thomas N Williams, David J Weatherall & Simon I Hay (2013). Global epidemiology of sickle haemoglobin in neonates: a contemporary geostatistical model-based map and population estimates, The Lancet, 381 (9861) 142-151. DOI: 10.1016/s0140-6736(12)61229-x

Roth E., Y Ueda, I Tellez, W Trager & R. Nagel (1978). Sickling rates of human AS red cells infected in vitro with Plasmodium falciparum malaria, Science, 202 (4368) 650-652. DOI: 10.1126/science.360396

Tapper, M. 1999. In the Blood: Sickle Cell Anemia and the Politics of Race. University of Pennsylvania Press. Google Books.

Tishkoff S.A., Alessia Ranciaro, Benjamin F Voight, Courtney C Babbitt, Jesse S Silverman, Kweli Powell, Holly M Mortensen, Jibril B Hirbo, Maha Osman & Muntaser Ibrahim & (2007). Convergent adaptation of human lactase persistence in Africa and Europe, Nature Genetics, 39 (1) 31-40. DOI: 10.1038/ng1946

Fucharoen S. (2012). The inherited diseases of haemoglobin are an emerging global health burden, Thalassemia Reports, 1 (2s) DOI: 10.4081/thal.2011.s2.e1

Wellems T.E. & Rick M. Fairhurst (2009). The impact of malaria parasitism: from corpuscles to communities, Journal of Clinical Investigation, 119 (9) 2496-2505. DOI: 10.1172/jci38307

Yi X., E. Huerta-Sanchez, X. Jin, Z. X. P. Cuo, J. E. Pool, X. Xu, H. Jiang, N. Vinckenbosch, T. S. Korneliussen & H. Zheng & (2010). Sequencing of 50 Human Exomes Reveals Adaptation to High Altitude, Science, 329 (5987) 75-78. DOI: 10.1126/science.1190371

[1] I would be inclined to forgive Wade for failing to perceive this rather technical distinction, except that he gives a rather detailed description of how hitchhiking works in chapter five of his book. So, apparently, although Wade does understand this subtle but important detail, he elected to omit this pesky point in favor of the much bolder pronouncement that “no less than 14% of the human genome … has changed under this recent evolutionary pressure [natural selection].” I suppose that a preference for attention-grabbing headlines over technical accuracy is what comes from years spent writing for the New York Time’s science section

[2] In preparing this section of the post, I was tempted to enumerate the many, many recent studies that suggest adaptation to different environments. In the interest of brevity, however, I decided to focus on the case studies that are included in Wade’s book, as all of them actually contradict his argument that natural selection has produced differences between races.

About these ads

5 comments on “A guide to the science and pseudoscience of A Troublesome Inheritance, part III: Has natural selection produced significant differences between races?

  1. Chris Smith says:

    Howdy Folks… The author here. Just a word about comments. Please be advised that comments are filtered for spam, and potentially spammy comments are individually moderated, but not by me. Moderation is handled by the NiB staff. Also note that if your comment violates NiB comment policy (which you can read here: http://nothinginbiology.org/policies/), your comment won’t be posted. For example, if your comment includes name-calling, NIB will not approve your comment.

  2. […] He also discusses in detail Wade’s repeated assertion that human evolution has been “recent, copious and regional”. […]

  3. […] week, at Nothing in Biology Makes Sense! Dismantling A Troublesome Inheritance, part III: has natural selection created differences between racial […]

  4. […] he’s not dismantling racist pseudoscience, Chris Smith studies the evolutionary ecology of species interactions. Willamette University sent […]

  5. I think that Steve Hsu raises a good point in noting that the methods used to identify differences between North and South Europeans in allele frequencies relating to height could ultimately test some of Wade’s speculations.

    The most controversial area is phenotypic differences in relation to cognitive abilities – something I understand Wade doesn’t go into in much detail. Davide Piffer has written some recent papers looking at allele differences in that respect for the handful of genes linked to date. Obviously until the genetic architecture for such traits is better understood people won’t be able to say with high confidence either way what the causal mechanisms are.

Comments are closed.