Continuous geographic structure is real, “discrete races” aren’t

| 64 Comments

Note: This topic is outside of my specialty, so it may be that I have missed some important points. I think I’ve got the basics correct, but this is a very complex topic. I will be interested in critical but constructive posts in the comments.

Update: required reading, which basically confirms my points I think:

Weiss, K. M. and J. C. Long (2009). “Non-Darwinian estimation: My ancestors, my genes’ ancestors.” Genome Research 19(5), 703-710.

Nievergelt, C. M., O. Libiger and N. J. Schork (2007). “Generalized Analysis of Molecular Variance.” PLoS Genetics 3(4), e51.

On Monday, Jerry Coyne at Why Evolution is True posted on “Are there human races?” While acknowledging the very bad history of the race concept in human history, and noting some of the problems with applying the concept to humans, Coyne concluded, basically, that the answer was yes, there are human races. While reviewing Jan Sapp’s piece which concluded that human races did not objectively exist, Coyne wrote:

As Sapp notes, and supports his conclusion throughout the review:

Although biologists and cultural anthropologists long supposed that human races–genetically distinct populations within the same species–have a true existence in nature, many social scientists and geneticists maintain today that there simply is no valid biological basis for the concept. The consensus among Western researchers today is that human races are sociocultural constructs.

Well, if that’s the consensus, I am an outlier. I do think that human races exist in the sense that biologists apply the term to animals, though I don’t think the genetic differences between those races are profound, nor do I think there is a finite and easily delimitable number of human races. Let me give my view as responses to a series of questions.

Unfortunately, while Coyne’s position is admirably clearly expressed, his reasons for accepting race don’t hold up. Some of his commentators have pointed out some of the problems in the comments. My own critique of his position is informed by a symposium I attended in San Francisco last year, “Is There Space for Race in Evolutionary Biology?”, and particularly by the arguments that speakers Massimo Pigliucci and Alan Templeton made there.

Here’s the conference information:

Is There Space for Race in Evolutionary Biology?

The 2011 Spring Conference for the Bay Area Biosystematists

Are human races geographical subspecies? How about distinct evolutionary lineages? What about ecotypes? Does it even make sense to divide humans into subspecies? Can there be biological races in other species even though there aren’t any in humans? This conference aims to answer these questions and more as it explores the question of whether race can be validly defined in the context of evolutionary biology. Come join us at the 2011 Spring Conference for the Bay Area Biosystematists! Please peruse this site to learn more about the conference.

Date: May 14, 2011 Time: 12:45-7pm Location: McLaren Conference Center, Room 250, University of San Francisco

Coyne asks a series of questions and gives answers:

1. What are races? Coyne answers:

“races of animals (also called “subspecies” or “ecotypes”) are morphologically distinguishable populations that live in allopatry (i.e. are geographically separated)”

2. Under that criterion, are there human races? Coyne answers:

Yes. As we all know, there are morphologically different groups of people who live in different areas, though those differences are blurring due to recent innovations in transportation that have led to more admixture between human groups.

3. How many human races are there? Answer:

That’s pretty much unanswerable, because human variation is nested in groups, for their ancestry, which is based on evolutionary differences, is nested in groups.

4. How different are the races genetically? Answer:

Not very different. […]

…but he goes on to cite the evidence of the results of clustering analyses that are often able use genetic data to place humans quite accurately into one population or another. I.e., the argument is basically that if there is enough signal in the genetics to place humans into one population or another, then races exist.

5. Why do these differences exist? Answer:

The short answer is, of course, evolution. The groups exist because human populations have an evolutionary history, and, like different species themselves, that ancestry leads to clustering and branching, though humans have a lot of genetic interchange between the branches!

He goes on to discuss the possible influence of natural and sexual selection on a few traits, and genetic drift and neutral evolution on the rest.

6. What are the implications of these differences? Coyne writes:

Not much. There are some medical implications.

…and goes on to (appropriately) dismiss certain pernicious ideas like the hard-to-kill one that IQ is genetically different between different races. However, he does leave open the possibility that a trait like intelligence could vary genetically between races, although it seems unlikely to him, he says there is no evidence supporting this, and in any event even a statistically-detectable difference between populations would be swamped by the within-population variability in such a trait. (Therefore, even in the event that such a population-level difference was found, one could not safely extrapolate from a race identification to a prediction about an individual.)

The critique

The key problems with the point of view that Coyne presents were made clear, at least to me, by Alan Templeton’s talk on this subject. I went to the talk in a quite skeptical mood, since Templeton is in my judgement on the wrong side of certain issues in phylogeography*. But Templeton’s points on race are simple and pretty convincing. See also:

Templeton, A. R. (1998). “Human Races: A Genetic and Evolutionary Perspective.” American Anthropologist 100(3), 632-650.

Race is generally used as a synonym for subspecies, which traditionally is a geographically circumscribed, genetically differentiated population. Sometimes traits show independent patterns of geographical variation such that some combination will distinguish most populations from all others. To avoid making “race” the equivalent of a local population, minimal thresholds of differentiation are imposed. Human “races” are below the thresholds used in other species, so valid traditional subspecies do not exist in humans. A “subspecies” can also be defined as a distinct evolutionary lineage within a species. Genetic surveys and the analyses of DNA haplotype trees show that human “races” are not distinct lineages, and that this is not due to recent admixture; human “races” are not and never were “pure.” Instead, human evolution has been and is characterized by many locally differentiated populations coexisting at any given time, but with sufficient genetic contact to make all of humanity a single lineage sharing a common evolutionary fate.

Taking Coyne’s points in order, I will summarize what I think Templeton would say, and add my own points.

1. What are races? Are human races “allopatric”? Templeton would deny that human populations “live in allopatry”. Humans are not, in fact, living in discrete, disconnected populations. Instead, humans are pretty continuously distributed from Africa all the way around the globe, and have been since glacial times (the tiny possible exceptions to this statement involve the recent colonization of remote islands e.g. by Polynesians, but this occurred only in the last few hundred-2000 years, a geological eyeblink). And genetic contact, i.e. interbreeding, seems to have been occurring between adjacent “populations” the whole time.

The overwhelming signal in human genetics seems to be isolation-by-distance – the larger the geographical distance between two samples, the greater the average genetic difference. But it is crucial to understand that this is not a confirmation of the idea of discrete races. If races were discrete, if you sampled along a transect, you would sample individuals from race A, A, A, A, and then suddenly switch to race B, B, B, B.

But what we typically see instead, if you sample along a transect, is that humans measured by some index of genetic difference along a transect would produce values with something like 0, 1, 2, 3, 4, 5, 6, 7. (I am making these numbers up, but imagine that they are a measurement of genetic difference from the first sample, i.e. the difference is 0 at the first sampling location, 1 at a sample 1000 km away, 2 at a sample 2000 km away, etc.)

This is a *crucial* distinction to get before we proceed, so I encourage readers to go back and re-read the previous paragraphs to get the idea into their head. Key concept: if you sample humans along a transect, the typical pattern is that genetic difference will increase in a continuous fashion with greater distance between samples, rather than in large discrete “steps” as you would expect with well-separated allopatric populations, or discrete “races”.

Here’s is Templeton’s depiction of this phenomenon:

Templeton_1999_AA_Fig5_geographic_distance_in_miles_vs_genetic_distance.png

2. Under that criterion, are there human races? Coyne answers “yes”, because we can see obvious differences between humans in different locations. But that, by itself, does not establish that the differences are discrete and categorical, which is what would have to be the case for ideas like “counting the number of races” and “distinguishing one race from the next” to make sense. Quite often, geneticists like Coyne seem to be being mislead by the fact that sampling is usually discrete. Usually, scientists go and get some samples from one region in Africa, some samples from one region in Europe, some from a region or two in Asia, etc. Before there was genetic sampling, there were similar issues in skull measurements, photographs, anthropologists’ monographs, etc.

When this form of sampling is done, of course it *looks like* the races are discrete, but what was actually discrete was your sampling. Templeton’s assertion is that if you had sampled on a grid, instead, and sampled all regions equally, you wouldn’t be able to draw discrete lines between races. The pattern would be that closer samples are more related, but this similarity would decline continuously with distance, and not in big discrete steps.

3. How many human races are there? Coyne says the question is “pretty much unanswerable”, and notes that the answers have varied from 3 to over 30. Coyne is correct to say the question is unanswerable, and the reason this is so is easily explained on Templeton’s view. But the unanswerability is quite a puzzle – or ought to be – if one is trying to argue that races are real. If you are going to assert that there is a rigorous and objective and scientific concept called “race”, but you can’t even explain how to objectively identify the races and thus count them, there’s a big problem somewhere.

For example, even though there are all kinds of definitions of “species”, and thus some variation in the count in the number of species, at least (a) the various species concepts are more-or-less objective, and (b) it happens that different concepts will often, although not always, give the same or similar answers. But, it is well-known that if “defining a species” is somewhat difficult and controversial, it is much worse for defining a subspecies, and it is close-to-hopeless for the ubiquitous concept of “population”. There are hundreds of papers on species concepts, but there are almost none published on “population concepts” (this, Millstein 2009, is one of the only ones; see also discussion of the paper). In practice, “populations” of organisms are basically either merely convenient geographic groupings determined subjectively by biologists for convenience or sampling purposes, very often geographically defined; or, they are convenient abstract entities useful for mathematical and computational modeling. But whether or not the “populations” you create for sampling and analysis purposes represent real discrete groups out there in the real world is an empirical question, and it looks like, in the case of humans, there is little basis for assuming discrete populations.

Finally, Coyne asserts at one point in his blogpost that races exist in humans much as subspecies exist in other species. But in terms of genetic divergence, at least, this is difficult to support. Templeton points out that the genetic divergence among living humans is much less than that typically observed between subspecies. This is not controversial or poorly known so I won’t argue this in detail. Templeton (1999) provides a graphical version:

Templeton_1999_AA_Fig1_Fst_for_humans_mammals_compared.png

(Note: see further discussion of criticism of Templeton (1998)’s Figure 1 at this comment.)

Or, we can compare the branch lengths of a tree of built of the aligned sequences from the mitochondria of humans and our ape relatives:

Gagneux_etal_1999_PNAS_mtDNA_African_hominids_Fig1_phylogeny.png

(From: Gagneux et al. 1999, PNAS; full-resolution figure) (Hat tip: Reed Cartwright)

Now, it is important to remember that the mitochondrial DNA sequence is really just one nonrecombining genetic marker, and it is possible for the mtDNA phylogeny to differ from the true species phylogeny (due to introgression) or to not reflect the pattern of diversity in the nuclear markers. Nuclear alleles coalesce more slowly, and recombination means that nuclear DNA will not have a simple branching phylogenetic history that can be easily represented by a tree, at least not within an interbreeding population. However, my sense of it is that in the great ape group (including humans), the mtDNA accurately represents the overall picture of phylogeny and genetic diversity.

5. Why do these differences exist? Coyne writes:

The short answer is, of course, evolution. The groups exist because human populations have an evolutionary history, and, like different species themselves, that ancestry leads to clustering and branching, though humans have a lot of genetic interchange between the branches!

Obviously we agree that selection, drift, etc. explain the evolution of observed diversity in the human species. But there are important questions here that Coyne doesn’t address. (1) How do we know that we’ve got discrete or even somewhat discrete “populations” which then can be said to have a distinct evolutionary history. Maybe we’ve just got one big interconnected population. This was addressed above. (2) How much interchange do you have to have to have before claims of clustering and branching break down? The evidence seems to indicate that the genetic differences between human “populations” in different regions are primarily a matter of differences in allele frequencies, but where most non-tiny populations have most of the alleles observed in the global population, just with varying frequencies. If this observation is true, this causes a problem for Coyne’s statement, because another way to say “the differences are differences in allele frequencies” is to say that “the differences are not fixed between populations”, and another way to say that is “coalescence has not occurred in these populations for these markers, either because of the lack of time or the presence of gene flow due to migration or both”. Yet another way to say this is to say that “for most loci, these populations are not reciprocally monophyletic”, and if you are a phylogeneticist, you known that this statements is equivalent to saying “your data are not appropriately represented by a single branching tree, instead you have a number of trees representing the individual histories of individual loci in the population”. Yet one more way of saying it is “you don’t have discrete populations, you are sampling within one genetically connected population.”

Coyne mentions clustering results, which have indeed impressed some people, and led to a minor resurgence in the race concept among geneticists in the last decade. It is true that, if you sample a region where people have mostly lived in or near the villages where they grew up for dozens of generations (e.g. parts of Europe, and e.g. not North America), you can take an individual’s genetic data and place their ancestry on the map with a fair degree of accuracy.

And it is also true that a computer algorithm, if given a number of samples from e.g. Africa, Europe, and China, could accurately place the samples into those 3 clusters, if told the number of clusters ahead of time. It is even true that newer algorithms can figure out the number of clusters in the data on their own, e.g. Structurama (full disclosure: authored by my advisor).

These sorts of observations have led a number of geneticists to speak of “Lewontin’s Fallacy” (e.g. Edwards 2003). The geneticist Richard Lewontin famously argued in 1972 that “about 85% of the total genetical variation is due to individual differences within populations and only 15% to differences between populations or ethnic groups” (Edwards 2003). Lewontin argued that this fact made the idea of racial classification irrelevant, since most variation was within groups instead of between them. You can see his point in an intuitive way by imagining that you have a sample of two people. Assuming the 85%/15% distribution of variation, for any one locus that varies between the two individuals, it is much more probable that the difference is just due to within-population variability than due to between-population differences.

Lewontin’s paper and argument have been cited thousands of times – for example, it was cited in just this fashion in the review by Sapp which Coyne is critiquing.

Edwards points out that the problem with Lewontin’s statement is that it is only true if you have only one locus to look at, or if you can safely assume that the variation at all loci is independent. It turns out that if you have, say, dozens or hundreds or thousands of loci to look at, and there is some population structure in the data, for example, population-level differences in allele frequencies, then by looking at all those loci, it becomes quite easy to identify the clusters and to accurately place individuals into them. Edwards describes how this works:

Consider two haploid populations each of size n. In population 1 the frequency of a gene, say “+” as opposed to “-“, at a single diallelic locus is p and in population 2 it is q, where p + q = 1. (The symmetry is deliberate.) Each population manifests simple binomial variability, and the overall variability is augmented by the difference in the means. The natural way to analyse this variability is the analysis of variance, from which it will be found that the ratio of the within-population sum of squares to the total sum of squares is simply 4pq. Taking p = 0.3 and q = 0.7, this ratio is 0.84; 84% of the variability is within-population, corresponding closely to Lewontin’s figure. The probability of misclassifying an individual based on his gene is p, in this case 0.3. The genes at a single locus are hardly informative about the population to which their bearer belongs.

Now suppose there are k similar loci, all with gene frequency p in population 1 and q in population 2. The ratio of the within-to-total variability is still 84% at each locus. The total number of “+” genes in an individual will be binomial with mean kp in population 1 and kq in population 2, with variance kpq in both cases. Continuing with the former gene frequencies and taking k=100 loci (say), the mean numbers are 30 and 70 respectively, with variances 21 and thus standard deviations of 4.58. With a difference between the means of 40 and a common standard deviation of less than 4.6, there is virtually no overlap between the distributions, and the probability of misclassification is infinitesimal, simply on the basis of counting the number of “+” genes. Fig. 1 shows how the probability falls off for up to 20 loci.

Edwards_2003_BioEssays_Lewontins_Fallacy_Fig1_probability_of_misclassification.png

This sort of information (obviously the between-population allele frequencies always differ between loci in real life, but this doesn’t matter much for the classification) is what lets the clustering algorithms find their clusters, and place people into them. Edwards concludes:

There is nothing wrong with Lewontin’s statistical analysis of variation, only with the belief that it is relevant to classification. It is not true that “racial classification is… of virtually no genetic or taxonomic significance”. It is not true, as Nature claimed, that “two random individuals from any one group are almost as different as any two random individuals from the entire world”, and it is not true, as the New Scientist claimed, that “two individuals are different because they are individuals, not because they belong to different races” and that “you can’t predict someone’s race by their genes”. Such statements might only be true if all the characters studied were independent, which they are not.

Lewontin used his analysis of variation to mount an unjustified assault on classification, which he deplored for social reasons. It was he who wrote “Indeed the whole history of the problem of genetic variation is a vivid illustration of the role that deeply embedded ideological assumptions play in determining scientific ‘truth’ and the direction of scientific inquiry”.(5) In a 1970 article Race and intelligence(11) he had earlier written “I shall try, in this article, to display Professor Jensen’s argument, to show how the structure of his argument is designed to make his point and to reveal what appear to be deeply-embedded assumptions derived from a particular world view, leading him to erroneous conclusions.”

A proper analysis of human data reveals a substantial amount of information about genetic differences. What use, if any, one makes of it is quite another matter. But it is a dangerous mistake to premise the moral equality of human beings on biological similarity because dissimilarity, once revealed, then becomes an argument for moral inequality. One is reminded of Fisher’s remark in Statistical Methods and Scientific Inference(12) “that the best causes tend to attract to their support the worst arguments, which seems to be equally true in the intellectual and in the moral sense.”

Edwards’s statistical argument is correct, but unfortunately – and ironically for someone criticizing Lewontin for not recognizing a mistake caused by a “deeply-embedded assumption” – Edwards is at least potentially being misled by his own deeply embedded assumption, as are others who argue along the lines of “clustering of humans based on genetics works, therefore different races exist.” Just because an analysis shows that the humans you sample fall into clusters does not necessarily mean that the clusters are out there in the real world in the human population. If the human population basically has continuous genetic variation, with genetic similarity decaying smoothly with distance, then, if your sampling was geographically clustered, your samples will show that structure.

I am not an expert in this area, but my sense of the matter from Templeton’s comments is that it is much more true to say that, at least for humans living basically in their ancestral locations, genetic difference between “populations” is essentially a smooth function of geographic distance. This is definitely “geographic structure” in the human population, but it is not evidence for discrete populations or races. And genetic clustering that is due to clustered sampling (done for convenience, funding reasons, political access reasons, and/or prior assumptions that localized sampling of a region represents that region, because you “already know” that a particular race inhabits that region) cannot be taken as evidence of real discrete structure in the human population.

It is possible that there is some discrete structure – due to language barriers or large migration events or whatever – but I would want to see this explicitly tested for rather than just assumed. And if it exists, it will be important to ask how much of genetic variation this explains, compared to the already major isolation-by-distance effect that we already know about and already know is quite strong.

A little more on clustering

One other point about clustering, this time with hierarchical clustering. Hierarchical clustering algorithms produce groups within groups, which are commonly represented by a tree structure very similar (superficially) to a phylogeny. You will sometimes see the argument that one can take the genetic distances between a bunch of samples of humans, throw them into a hierarchical clustering algorithm, and get a tree out, and therefore we can safely think of human races as subspecies within our species. This matches up superficially with a lot of our background conceptual model of how species can be placed as groups within groups on a phylogeny, and thus tends to reinforce the “races are real” idea in some peoples’ heads.

However, this argumentation is at least dubious. First, clustering methods taken as a phylogenetic method are a form of “distance method”, and distance methods are well-known in phylogenetics to be the crudest sort of phylogenetic method and the one most prone to errors; unless some pretty specific assumptions are met (e.g. molecular clock) they can produce very bad errors. Basically, distance methods rely on the pairwise Euclidean distances between samples in your data space. You can imagine this as the “as the bird flies” distance. However, the distance that we are really interested when we produce a phylogeny is not the “as the bird flies”, Euclidean distance between our samples; instead, we are most interested in the path distance, i.e., the actual mutational steps (or character-change steps) that led from a common ancestor, up each branch, to the tips that we observe. This form of distance is known as a “Manhattan distance”, named after the square street grid in downtown New York. In New York City, you can’t get from A to B by going “as the crow flies”, you have to drive along the street grid, avoiding one-way streets, construction, traffic jams, etc. Only in idea situations will the actual traveled path distance equal or closely approximate or correlate linearly with the as-the-crow-flies distance. Phylogenetic methods are thus really about finding the best paths (branching patterns) between the observed samples, according to some criterion (e.g. parsimony, maximum likelihood, posterior probability).

Therefore, clustering methods are not phylogenetic methods, except in special situations.

Second, a hierarchical clustering method will produce hierarchical clustering from almost any but the most artificially-constructed distance data. If there is true hierarchical structure, the method can find it, but some sort of clustering will be produced, even if it’s poorly supported and changes dramatically under e.g. resampling of the data. And there can be many sources of any clustering that is observed. I bet that you could:

* lay down a latitude/longitude grid on the earth,

* take a “sample point” at every intersection point between latitude and longitude lines (done at, say, 1 degree or 5 degree intervals)

* exclude points that fell in the ocean

* calculate the great circle geographic distance between all points

* input this into a hierarchical clustering algorithm

…and you would get out a nice treelike diagram that would show each of the major continents as clusters, probably with linked continents forming groups of groups.

This would be successful hierarchical clustering, for sure, but it obviously wouldn’t be telling you about the evolutionary connectedness of the continents or anything like that. It’s a summary of the geographical distance matrix between the samples.

The interesting question is – what if you had the samples to calculate the genetic distances between the (indigenous) humans sampled at each of these points. Would you get a clustering scheme all that much different than the one you would get by clustering pure geographic distance? Probably there would be some differences, since we know that humans started in Africa and genetic diversity is highest there – but would it be a big effect? Or would the “clustering” that is observed in the genetic distances primarily just be the effect of isolation by distance and the fact that humans live on land, which is unevenly distributed around the globe? This might verify “race” in a very vague way, but it would be a long way from the traditional concepts that most people have in their head.

Lastly, we reach point 6.

6. What are the implications of these differences? Coyne writes:

Not much. There are some medical implications. […]

Here, interestingly, Templeton might actually agree, at least for the United States. Although it is illegitimate to do discrete sampling of continuous variation, and then conclude that discrete variation exists in reality, it is important to note that in some cases, “discrete sampling” in history has played a role in producing modern populations. For example, the population of the U.S. is not a random sample of the global human population. For historical reasons, in the formation of the U.S. population there really were major contributions from discrete parts of the globe – Europe, West Africa, East Asia, etc. For this reason, if you were to examine a few hundred randomly sampled Americans, you would indeed find detectable discrete genetic variation! In this (very limited) sense, “race” is physically real, but that physical reality was completely constructed by contingent cultural events in recent history!

In certain cases, this might even be medically relevant, for example if certain genotypes are common in a certain population, and these genotypes react especially well or especially poorly to a particular drug. Certainly you would want to know this if you were a doctor prescribing medication to a patient, and government agencies like the FDA will want to know it when they are approving drugs for sale or for insurance coverage. And certainly, it seems arbitrary and unfair to do drug tests on one group (say, white males) and then assume that the results will apply with equal validity to everyone else.

Finally, the fact that there is some geographical structure (whether discrete or continuous) in the human population, or in the weird sample of the human population that we have in the U.S., is important to take into account for statistical reasons. Many studies are attempting to correlate genetic traits with various diseases and conditions. But these studies can be badly misled if there is correlation in the genetic data that is due to, say, geographic ancestry that is unrecognized. This is a vast field, and pretty much represents a huge job market for evolutionary biologists and population geneticists who decades ago would have puttered around putting equations on chalkboard, but who are now crucial components of any well-done study involving genetic data from populations.

Those are some of the arguments favoring taking “race” into account in medicine and medical research. However, as you can imagine, there are all kinds of difficulties and dangers. A big one is that some statistical effect that is true in one situation – say, people who identify as African American have a higher risk of condition X – will be illegitimately extrapolated to other people who fall into doctors’ and the public’s cultural definition of African American. What if one’s ancestry is from southern Africa, but the genetic trait that correlates with the condition derives from West Africa? The statistical generalization made in some study may not apply to you.

This post is long enough, but hopefully readers can see that the topic of race in the human species is a complicated one, fraught with danger even from a purely scientific point of view, even before we get to the even more hazardous arenas of culture and politics. I think that it should be clear that one cannot give the question “Do races exist?” any kind of simple “yes” answer in humans. One can say “yes” to the question, “Is there geographical structure in human genetics?”, but hopefully I have shown that this is a long way from establishing that there is any kind of discrete genetic structure in the human population at large, i.e. discrete “racial groups” which could be identified and counted. It is possible that with ever more data, some shadow of this idea may be statistically supportable, but it seems unlikely that it will compare in strength to the strong overall pattern of continuous genetic variation following an isolation-by-distance model.

Conclusion

I’ll let the conclusion of Templeton (1998) speak for itself:

Conclusions

The genetic data are consistently and strongly informative about human races. Humans show only modest levels of differentiation among populations when compared to other large-bodied mammals, and this level of differentiation is well below the usual threshold used to identify subspecies (races) in nonhuman species. Hence, human races do not exist under the traditional concept of a subspecies as being a geographically circumscribed population showing sharp genetic differentiation. A more modem definition of race is that of a distinct evolutionary lineage within a species. The genetic evidence strongly rejects the existence of distinct evolutionary lineages within humans. The widespread representation of human “races” as branches on an intraspecific population tree is genetically indefensible and biologically misleading, even when the ancestral node is presented as being at 100,000 years ago.

Attempts to salvage the idea of human “races” as evolutionary lineages by invoking greater racial purity in the past followed by admixture events are unsuccessful and falsified by multilocus comparisons of geographical concordance and by haplotype analyses. Instead, all of the genetic evidence shows that there never was a split or separation of the “races’” or between Africans and Eurasians. Recent human evolution has been characterized by both population range expansions (with perhaps some local replacements but no global replacement within the last 100,000 years) and recurrent genetic interchange. The 100,000 years ago “divergence time” between Eurasians and Africans that is commonly found in the recent literature is really only an “effective divergence time” in sensu Nei and Roychoudhury (1974, 1982). Since no split occurred between Africans and Eurasians, it is meaningless to assign a date to an “event” that never happened. Instead, the effective divergence time measures the amount of restricted gene flow among the populations (Slatkin 1991).

Because of the extensive evidence for genetic interchange through population movements and recurrent gene flow going back at least hundreds of thousands of years ago, there is only one evolutionary lineage of humanity and there are no subspecies or races under either the traditional or phylogenetic definitions. Human evolution and population structure have been and are characterized by many locally differentiated populations coexisting at any given time, but with sufficient genetic contact to make all of humanity a single lineage sharing a common, long-term evolutionary fate.

My recommendation would be to ditch the discussion of “human races” in science wherever possible (it may be unavoidable in certain fields, e.g. medicine, due to its cultural reality), particularly in genetics, and instead go with “geographic ancestry” instead. Some of us humans will trace back to a small region somewhere, others of us will be fusions from two or five or six continents. Coyne’s post argued that races exist, but were becoming less distinct due to interbreeding. I would say that continuous geographic structure exists, but that it is becoming weaker as the human population becomes more and more panmictic.

Note: After I wrote the above, I noticed that Coyne has a followup post up that cites the famous “map of Europe reflected in genetics” (paraphrasing) figure from Novembre et al. (2008). He notes the evidence for the isolation-by-distance effect, and concludes that there are no races in Europe, at least:

In other words, genetically closer populations are more genetically similar, as expected if individuals tend to mate with other individuals from the same country, and close by. This is an “isolation by distance” model: genetic similarity falls off gradually with distance. As the authors note, this does not support the existence of “discrete, well-differentiated populations,” i.e., there are no races. None are expected in such a small area, particularly because biological “races” are those populations that (at least at one time) were geographically isolated and genetically differentiated. That geographical isolation never happened in Europe.

[…]

As I said, this doesn’t show that there are discrete “races” in Europe, and I don’t think there are obviously discrete “races” anywhere these days, though there is large-scale genetic differentiation among worldwide population suggesting that such races once existed as relatively discrete and geographically isolated populations. The discreteness that once existed, or so I think, is now blurring out as transportation and migration are beginning to mix the discrete groups into not a melting pot, but sort of a lumpy pudding of humanity.

What is clear is that, with considerable accuracy, you can diagnose an individual’s geographic origin from his genes. Nearly everyone’s DNA contains reliable information about their recent and ancient past. We are not all genetically alike. If we were, you couldn’t do studies like the one of Novembre et al. But neither are we radically different genetically, for if we were, you wouldn’t need hundreds of thousands of genes for such accurate predictions.

So Coyne is actually approaching the Templeton view of human genetic diversity pretty rapidly, although he is still buying into “relatively discrete” races at a larger scale (what are they then?), which I would disagree with.

References

Edwards, A. W. F. (2003). “Human genetic diversity: Lewontin’s fallacy.” BioEssays 25(8), 798-801.

Gagneux, P., C. Wills, U. Gerloff, D. Tautz, P. A. Morin, C. Boesch, B. Fruth, G. Hohmann, O. A. Ryder and D. S. Woodruff (1999). “Mitochondrial sequences show diverse evolutionary histories of African hominoids.” Proceedings of the National Academy of Sciences 96(9), 5077-5082.

Lewontin RC. The apportionment of human diversity. In: Dobzhansky T, Hecht MK, Steere WC, editors. Evolutionary Biology 6. New York: Appleton-Century-Crofts. 1972. p 381-398.

Millstein, R. L. (2009). “Populations as Individuals.” Biological Theory 4(3), 267-273.

Templeton, A. R. (1998). “Human Races: A Genetic and Evolutionary Perspective.” American Anthropologist 100(3), 632-650.

Notes

* For example (1) he promotes a method that seems to be subjective and to not work (it’s called Nested Clade Analysis, it’s not about nested clades in the traditional phylogenetic/macroevolutionary senses of those words, and it would take awhile to explain…but see “Why does a method that fails continue to be used?” by Lacy Knowles for the basics, and Google Scholar for the ongoing discussion.) and (2) he doesn’t seem to get Bayesian methods and his one-man war against ABC (Approximate Bayesian Computation) consists primarily of misunderstandings of Bayesian logic (although ABC does have various limitations and problems, they aren’t primarily the ones he identifies; see e.g. “Invalid arguments against ABC: Reply to AR Templeton” by Csillery et al.).

However, NCA is Templeton’s pet method and was popular for a long time (thousands of citations), and Bayesian logic can be difficult even for very smart professionals, especially if they have been raised thinking of frequentism as the only way to think about statistics, and Templeton has vast experience in human population genetics and popgen applied to medicine, so I don’t think his opinions on broader, simpler issues can be dismissed because of the fight over NCA/ABC.

I might as well add that my specialty is evolutionary biogeography at the macroevolutionary level of large number of species related by a phylogeny, not phylogeography, which operates primarily at the level of population genetics of one species or a few closely-related species/populations, so my opinions in this arena are still developing as I learn more about it.

** Another note: re-reading Templeton (1998) after some time just now, I found rather more Nested Clade Analysis in there than I remembered. I don’t think this affects the parts of his conclusions that I quoted, except perhaps for his skepticism of the 100,000 year “divergence time” for Africans vs. non-Africans and his endorsement of an alternative model (derived from his “trellis model”).

Templeton in 1998 is in part lobbying for the “trellis” model, where all humans are connected by gene flow and always have been, all the way back to Homo erectus. He makes the modern isolation-by-distance observation a part of the evidence for this, but I don’t see any reason to endorse this. We could have a fairly standard Recent Out-of-Africa model (adding minor interbreeding with Neandertals etc.) and still get the isolation-by-distance effect as long as humans have the behavior of breeding with nearby tribes and thus having high gene flow but mostly with adjacent groups.

Other links of interest:

http://en.wikipedia.org/wiki/%22Hum[…]fic_paper%29

64 Comments

As far as I can tell Coyne has no point here; he claims that human races exist but can’t say how many there are (and therefore, by extension, if he can’t identify what the races are in order to count them, he can’t say who belongs to what “race”).

Meanwhile, there is unequivocally a sufficient association of certain medical issues with certain socially-defined ethnic identities, to warrant recognition of a patient’s self-identified ethnicity in some circumstances.

Many associations are the result of environmental factors. For example, “Mormon” isn’t an ethnic group, but it is a group of people who have a low rate of pathologies related to excess consumption of ethanol or use of cigarettes. However, if an originally Mormon person takes up heavy drinking and smoking, that characteristic disappears. This is an example of how membership in a culturally defined group can be related to risk of certain diseases, without any genetic input.

On the other hand, sometimes there is an association which may have something to do with genes. People of East Asian and Amerindian descent have different rates of certain types of lymphoma/leukemia (fortunately these particular diseases are extremely rare; their incidence would not have significant impact on overall life expectancy of any population, but people get them and they need to be diagnosed and treated). As it happens, knowing the patient’s ethnic group has absolutely nothing to do with the process of diagnosing these disorders, but if this association holds up, it might lead to a clue that would help us to better understand that pathogenesis of these disorders. Note that risk of these rare disorders is not increased in one “race” as culturally defined. The disorders discussed in the references below are also strongly associated with viruses, HTLV-1 and EBV, respectively. Other environmental factors can’t be ruled out. Nevertheless, they occur in multiple diverse environments, but the increased (although still very low) risk in East Asian and Amerindian populations seems to hold up.

http://www.ncbi.nlm.nih.gov/pubmed/21205073

http://www.ncbi.nlm.nih.gov/pubmed/21738302

http://www.ncbi.nlm.nih.gov/pubmed/21510468

Conclusion - “Race” is a social construct based on clustering of a few superficial characteristics that often lumps together populations that aren’t even closely related. However, pathology can associate with culturally defined ethnicity, probably usually for environmental reasons, but sometimes perhaps for genetic reasons.

Any of us, who took Biological Anthropology at the University of Michigan, from the 60’s until a year or so ago, had the chance to take C. Loring Brace’s course on race, which made most, if not all these points. The faculty in my time there (mid-late 1970’s) had little good to say about the concept of human races at all. It was hard to cling to ideas that races were “real” after hearing Conrad Kottak talk about racial taxonomy in Brazil (where he had done fieldwork).

IMHO, this is old news, but I suppose race is one of those concept that will take a loooong time to die its well deserved death.

dpr

D P Robin said:

Any of us, who took Biological Anthropology at the University of Michigan, from the 60’s until a year or so ago, had the chance to take C. Loring Brace’s course on race, which made most, if not all these points. The faculty in my time there (mid-late 1970’s) had little good to say about the concept of human races at all. It was hard to cling to ideas that races were “real” after hearing Conrad Kottak talk about racial taxonomy in Brazil (where he had done fieldwork).

IMHO, this is old news, but I suppose race is one of those concept that will take a loooong time to die its well deserved death.

dpr

I thought it was old news also, but then I saw Jerry Coyne, a famous population geneticist and current head of the Society for the Study of Evolution making these arguments.

Uh, what is NUMT in the big chart above?

Summary: the boundary between “races” is way fuzzier even than the boundary between closely related species.

Also the differences between “races” is often swamped by the differences within each of them.

I suppose that a long long time ago (in a galaxy far far away?), Europe and Asia may have had groups more or less distinct from each other or the ones in Africa (though I wouldn’t count on it), but as the above points out, that ain’t the case today. (Oh, and “African” is certainly not a distinct group!)

Henry

Following the death of Christopher Hitchens, Coyne posted an impassioned defense of virulent misogyny. That’s when I stopped reading him. Unfortunately, in the early 21st century, “Conservatism” trumps science, atheism, and rationalism. I wish this weren’t true.

It’s a shame, too, because he’s potentially a valuable person, and not a liability. But these are the times we live in.

Nick Matzke said:

D P Robin said:

Any of us, who took Biological Anthropology at the University of Michigan, from the 60’s until a year or so ago, had the chance to take C. Loring Brace’s course on race, which made most, if not all these points. The faculty in my time there (mid-late 1970’s) had little good to say about the concept of human races at all. It was hard to cling to ideas that races were “real” after hearing Conrad Kottak talk about racial taxonomy in Brazil (where he had done fieldwork).

IMHO, this is old news, but I suppose race is one of those concept that will take a loooong time to die its well deserved death.

dpr

I thought it was old news also, but then I saw Jerry Coyne, a famous population geneticist and current head of the Society for the Study of Evolution making these arguments.

YEC here. The bible says the people were one and then they divided up and covered the earth. Is there races? Well if merely being very segregated people groups and this a good long time equals race then there is races. It doesn’t matter how different people become and small or great differences are just manifestations of being segregated peoples. For example I am of the great tribes of German and Celt. In no way am i biologically close to the slavic peoples. We were segregated at babel or at least everyone must agree since the language changes. Yet I look exactly like this people. THis just because of the same adaptations due to the Northern areas we live in. Likewise the great differences in men are from triggers in our bodies in reaction to needs in the world. A problem here is that evolution is teaching small steps leading to big changes forces conclusions of great genetic segregation. Yet in fact its just quick physical adaption that affects or doesn’t segregated populations in like or unlike ways.

They are trying to say there is a white Adam and Eve and then the people segregated into different languages. instead there was first the language segregation in Southern areas and then these segregated peoples moved into northern areas and equally adapted for like needs but nowhere near each other. JUst like animals in the north who all get white fur.

By the way its brought again, as a careful option, that iQ is different. some smarter then others upon birth. This is impossible for many reasons yet shows that giving in a little on this will lead to whatever anyone thinks. Yes different people and different peoples are “intelligent” in different degrees but this is just about time and place. all babies are equal in bringing no intelligence with them. its all learnt after birth with results relative to motivation and whats within reach.

Evolutionary presumptions are what confuses matters.

Robert Byers said:

YEC here. The bible says the people were one and then they divided up and covered the earth.

It does not matter what a 2000 year old book says. Only what can be discovered in the present reality we all live in.

Is there races? Well if merely being very segregated people groups and this a good long time equals race then there is races. It doesn’t matter how different people become and small or great differences are just manifestations of being segregated peoples. For example I am of the great tribes of German and Celt. In no way am i biologically close to the slavic peoples. We were segregated at babel or at least everyone must agree since the language changes. Yet I look exactly like this people. THis just because of the same adaptations due to the Northern areas we live in. Likewise the great differences in men are from triggers in our bodies in reaction to needs in the world. A problem here is that evolution is teaching small steps leading to big changes forces conclusions of great genetic segregation. Yet in fact its just quick physical adaption that affects or doesn’t segregated populations in like or unlike ways.

They are trying to say there is a white Adam and Eve and then the people segregated into different languages. instead there was first the language segregation in Southern areas and then these segregated peoples moved into northern areas and equally adapted for like needs but nowhere near each other. JUst like animals in the north who all get white fur.

By the way its brought again, as a careful option, that iQ is different. some smarter then others upon birth. This is impossible for many reasons yet shows that giving in a little on this will lead to whatever anyone thinks. Yes different people and different peoples are “intelligent” in different degrees but this is just about time and place. all babies are equal in bringing no intelligence with them. its all learnt after birth with results relative to motivation and whats within reach.

You clearly have never studied human brains and how they function.

Evolutionary presumptions are what confuses matters.

Once again, you have lied outright. That’s about all you do, it seems.

There are times when Byers is embarrassing, and times when he’s painful. This is two of those times.

@diogeneslamp0,

NUMT is a sequence of mitochondrial origin, that was copied into the nuclear genome. The tree is based on a mitochondrial genome segment.

The relevant part from the paper, “We examined part of hypervariable CR1, corresponding to the human mtDNA nucleotide positions 16,053–16,465 (16), from 1,158 unique haplotypes, almost all (,1000) of known geographic provenance. These include 83 previously unpublished haplotypes representing 70 chimpanzees and 13 bonobos and 1,070 published sequences from GenBank (811 humans from around the world, 26 gorillas, 11 bonobos, and 222 chimpanzees). In addition, we included one Neanderthal (17) and one human nuclear sequence of mitochondrial origin (numt) involving a CR1 sequence that became inserted into chromosome 11 sometime in the past (18). Haplotypes compared ranged in length from 300–415 bp except for three comprising 270–298 bp that were included to capture the full extent of known variability. Three published orangutan sequences representing both subspecies were used as an outgroup for analytical purposes (19, 20). GenBank accession nos. of all sequences used, with their 35 specific source references and our alignments, are available from P.G.

-The Other Jim

Byers and Coyne on the same page?

How about the Aboriginal Australians? As far as I know they were genetically isolated for around 50,000 years. Could they form a race as you’re defining it?

Robert Byers said:

YEC here. The bible says the people were one and then they divided up and covered the earth. Is there races? Well if merely being very segregated people groups and this a good long time equals race then there is races. It doesn’t matter how different people become and small or great differences are just manifestations of being segregated peoples. For example I am of the great tribes of German and Celt. In no way am i biologically close to the slavic peoples. We were segregated at babel or at least everyone must agree since the language changes. Yet I look exactly like this people. THis just because of the same adaptations due to the Northern areas we live in. Likewise the great differences in men are from triggers in our bodies in reaction to needs in the world. A problem here is that evolution is teaching small steps leading to big changes forces conclusions of great genetic segregation. Yet in fact its just quick physical adaption that affects or doesn’t segregated populations in like or unlike ways.

They are trying to say there is a white Adam and Eve and then the people segregated into different languages. instead there was first the language segregation in Southern areas and then these segregated peoples moved into northern areas and equally adapted for like needs but nowhere near each other. JUst like animals in the north who all get white fur.

By the way its brought again, as a careful option, that iQ is different. some smarter then others upon birth. This is impossible for many reasons yet shows that giving in a little on this will lead to whatever anyone thinks. Yes different people and different peoples are “intelligent” in different degrees but this is just about time and place. all babies are equal in bringing no intelligence with them. its all learnt after birth with results relative to motivation and whats within reach.

Evolutionary presumptions are what confuses matters.

So that would be a no, you can’t deal with the evidence at all. Too bad for you Robert.

As for Adam and Eve, they were obviously black, but that’s OK, undoubtedly Jesus was too. Get a clue man.

BIblical presuppositions are what confuses you Robert, get rid of them and deal with reality.

I suggest an immediate dump to the bathroom wall for Robert and his fantasies. Why derail an otherwise intelligent conversation?

It seems we are rehashing Darwin, yet again. OOS 6th edition, chapter 2, section 3 (titled “Doubtful Species”):

How many of the birds and insects in North America and Europe, which differ very slightly from each other, have been ranked by one eminent naturalist as undoubted species, and by another as varieties, or, as they are often called, geographical races!

He even covered Nick’s sampling issue!

After this discussion, the result of so much labour, he [A. de Candolle] emphatically remarks: “They are mistaken, who repeat that the greater part of our species [oaks] are clearly limited, and that the doubtful species are in a feeble minority. This seemed to be true, so long as a genus was imperfectly known, and its species were founded upon a few specimens, that is to say, were provisional. Just as we come to know them better, intermediate forms flow in, and doubts as to specific limits augment.”

Maybe its time to ressurrect Darwin’s term ‘doubtful species,’ and apply it to ourselves.

Robert Byers said: …iQ is different.

Would you happen to know your IQ, Robert? Do you know if it’s positive or negative?

Robert Byers said:

YEC here… Is there races? …there is races.

EV here. Is the creationists learning? No.

EV here.

Thanking Masked Panda for definition of NUMT in above graph. But why is NUMT such an outlier?

Among the humans, who are the genetic outliers?

Khoisan? Nuristani? Melanesians?

The cast of Jersey Shore?

Dave Luckett said:

There are times when Byers is embarrassing, and times when he’s painful. This is two of those times.

Agreed. But he’s utterly pathetic without any hope for redemption. What more do you expect?

diogeneslamp0 said: Thanking Masked Panda for definition of NUMT in above graph. But why is NUMT such an outlier?

It’s some combination of two possibilities. Nuclear mtDNAs (NUMTs):

(a) are basically junk DNA, at least usually, and so experience higher substitution rates than a sequence under stabilizing selection would. (the functioning genes in the mtDNA are under stabilizing selection, for instance)

(b) It’s quite possible that that particular NUMT originated in the human lineage back in the Homo erectus days. For example, perhaps some mitochondrion in some germline cell was damaged, its DNA was floating around, and some of it got spliced into the nucleus by mistake by some DNA repair enzymes. Then, by chance, this got fixed in the population, or at least became frequent enough to get sampled.

And actually, (b) is a pretty interesting idea if you think about it. If we had enough of these, we’d have some “fossil” genetic information about ancestral populations – whereas with actual hominid fossils, we will probably never be able to get DNA from a Homo erectus, since the time limit on DNA is about 100,000 years even in the best conditions.

See this paper I was recently on, we discovered that some mtDNAs in domestic horses that had been published in genbank seemed to stick out against the background of a bunch of sequences that we had. One possibility is that these were NUMTs that had accidentally been sequenced and were mistakenly thought to be true mtDNAs. When you grind up some cells and try to amplify your target mtDNA sequence using PCR, there is no guarantee you won’t accidentally amplify a NUMT instead, without careful double-checking. (Ditto for having a computer reconstruct a sequence from a library of short sequences produced by shotgun sequencing. The computer could accidentally stick some NUMT sequence into your mtDNA sequence.)

Lippold, S., N. J. Matzke, M. Reissmann and M. Hofreiter (2011). “Whole mitochondrial genome sequencing of domestic horses reveals incorporation of extensive wild horse diversity during domestication.” BMC Evolutionary Biology 11(1), 328.

Among the humans, who are the genetic outliers?

Khoisan? Nuristani? Melanesians?

The cast of Jersey Shore?

If the argument I presented in the main post is correct, this is a meaningless question. The biggest overall genetic distance, measured across many loci, will be between samples that are (ancestrally, pre-1600s) geographically distant.

In just the mtDNA sequence, the biggest diversity and deepest branches are in Africa, and this is also true for many individual loci in the nuclear DNA (although not all, e.g. the “Neandertal markers” are deeply-branching loci in Eurasians). But this cannot be extrapolated to make statements about populations and genetic distance, because of interbreeding. The mtDNA tree will not line up with the tree from nuclear locus #1, which won’t line up with nuclear locus #2, etc. This all washes out when averaged across many loci, and you get the observation that genetic similarity correlates with geographic distance.

Byers and Coyne on the same page?

No. My disagreement with Coyne is fairly minor in the grand scheme of things. Byers is on another planet.

howard.peirce said:

Following the death of Christopher Hitchens, Coyne posted an impassioned defense of virulent misogyny. That’s when I stopped reading him. Unfortunately, in the early 21st century, “Conservatism” trumps science, atheism, and rationalism.

Where did he post that, and what was his logic?

Um, should we read Coyne out of the movement now? I’m still steamed about that goddamn peppered moth thing. To think that the IDologues were citing COYNE in support of their mottephobic conspiracy theory!

I’m reading Coyne’s “Why Evolution Is True”, and noticed outrageous errors. At one point he writes that some woodpecker’s tongues are anchored in their nostrils! That’s a creationist myth. He can’t double-check that on TalkOrigins? You realize TalkOrigins is far more accurate than a tree-killin book?

Please ignore format of the post above.

howard.peirce said:

Following the death of Christopher Hitchens, Coyne posted an impassioned defense of virulent misogyny. That’s when I stopped reading him. Unfortunately, in the early 21st century, “Conservatism” trumps science, atheism, and rationalism. I wish this weren’t true.

Where did he post that, and what was his logic?

Um, should we read Coyne out of the movement now? I’m still steamed about that goddamn peppered moth thing. To think that the IDologues were citing COYNE in support of their mottephobic conspiracy theory!

I’m reading Coyne’s “Why Evolution Is True”, and noticed outrageous errors. At one point he writes that some woodpecker’s tongues are anchored in their nostrils! That’s a creationist myth. He can’t double-check that on TalkOrigins? You realize TalkOrigins is far more accurate than a tree-killin book?

Matt,

I personally believe that I can pretty much figure out if someone traces his ancestors to Sweden, India, East Africa or Japan, or some place in the general region. Rednecks also believe this. Moreover, the redneck will see that you and your science are barking mad when you claim, using mumbo-jumbo technicalities, that human races do not exist. Every time biology or medicine finds another population-specific trait, racists are vindicated, and science falsified.

I personally believe that it is necessary for biology to recognize that human races exist, with the caveat that they are neither sharply defined genetically or biologically, that the genetic differences that exist are small and of recent origin (abt. 50000 years; 2000 generations), and that whatever population differences that may exist can be usefully combined through miscegenation. In any case, Edwards is right – denying the geography of genetics is not a solution.

Previous “Masked Panda” is W. Benson

Since you people in general know a lot more then I do, I don’t normally get to contribute.

However in this case I think I have something visual that can help get the point across.

On second thought this is also a question which I will get to in a moment. But first what I think:

If you look at geographical distribution of a genetic traits using a map, you often will notice that the distribution doesn’t really line up with the boarders of races. For example this link http://en.wikipedia.org/wiki/File:W[…]olor_map.jpg

Shows the distribution of blue eyes in Europe. There is not any distinction that I can tell separating Germanic from Slavic or Francs for example. In fact the distribution seems to ignore most human ideas for territories and racial ancestry.

I am sure someone here can correct me if I am wrong but I don’t recall any similar map that does respect such boundaries. My question is am I correct that this pattern is more or less repeated for any particular genetic trait?

Incidentally looking at the resident troll, I have both Slavic and Germanic ancestry and I got my blue eyes from both sides. The groups aren’t nearly as segregated as it imagines. (And the diagram shows)

Matt,

I personally believe that I can pretty much figure out if someone traces his ancestors to Sweden, India, East Africa or Japan, or some place in the general region. Rednecks also believe this.

I doubt if you can do this as accurately as you seem to think; nevertheless, it is irrelevant even if you can.

No-one here is denying the obvious fact that superficial traits like skin/hair/eye pigmentation and some others cluster geographically. It is very clearly explained in the original post why this is different from biological recognition of “races”.

Moreover, the redneck will see that you and your science are barking mad when you claim, using mumbo-jumbo technicalities, that human races do not exist.

No-one has made this claim either. Of course the cultural construct “race” exists. Of course, in US society, it is related to pigmentation of skin, hair texture, and some other features.

The term “race” is used differently in biology. The point here is that culturally recognized “races” of humans are not the same thing as biological races.

Every time biology or medicine finds another population-specific trait, racists are vindicated, and science falsified.

Racism is the subjective ethical decision to deliberately treat some people badly or unfairly because of their perceived ethnic origin. I personally prefer the term “ethnic bigotry” for clarity (since it is common for people to be bigoted against others whom they define as belonging to a different ethnic group, even if the two groups cannot be distinguished based on physical appearance), although I sometimes use “racism” because it is a more powerful word.

To repeat for emphasis, the decision of whether or not to be a racist is purely a subjective ethical decision. For example, people with trisomy 21 unequivocally are genetically different from other humans in a significant way, and are disproportionately likely to have certain problems due to genetics. Yet I oppose treating them badly or unfairly. In an opposite example, various closely related Scandinavian populations who barely spoke different dialects, and were indistinguishable to outsiders, have been quite bigoted against each other in the past. Science cannot tell you whether it is “good” or “bad” to treat people unfairly.

As it happens, racists often do make exaggerated claims that are falsified by science.

I noted above that certain hematopoietic malignancies have different rates of incidence in different populations. This does not conflict with Nick Matzke’s point about race.

I personally believe that it is necessary for biology to recognize that human races exist, with the caveat that they are neither sharply defined genetically or biologically, that the genetic differences that exist are small and of recent origin (abt. 50000 years; 2000 generations),

If you agree with that, then what are you arguing about?

and that whatever population differences that may exist can be usefully combined through miscegenation.

No doubt you are unaware of the obnoxious historical associations, and thus, potentially offensive nature, of the word “miscegenation”. If it is not your objective to deliberately offend, or to send coded signals of approval of racism - and hopefully it is not - I recommend that you use a synonym in the future.

In any case, Edwards is right – denying the geography of genetics is not a solution.

Luckily, no-one is doing that at all, nor denying the existence of the cultural construct “race”. The point is that genetic differences between human populations do rise to the level of designation of different biological races.

***Human “races” are below the thresholds used in other species, so valid traditional subspecies do not exist in humans. A “subspecies” can also be defined as a distinct evolutionary lineage within a species. ***

Templeton unfortunately appears to have based this idea on a misreading of Smith et als’s 1997 article from Herpetological Review entitled “Subspecies and Classification”. Templeton suggested that there needed to be a FST of at least 25%-30%. However, Smith et al were not referring to FST. They were referring to the 75% rule for identifying sub species.

Incredibly few people have bothered to go back and read Smith’s paper!

*** Key concept: if you sample humans along a transect, the typical pattern is that genetic difference will increase in a continuous fashion with greater distance between samples, rather than in large discrete “steps” as you would expect with well-separated allopatric populations, or discrete “races”.***

If you read Rosenberg’s 2005 paper on Clines, Clusters, and the Effect of Study Design on the Inference of Human Population Structure there are in fact discrete clusters:

“That is, for the five clusters with K = 5 in Figure 2 of the present study and in Figure 1 of [3]—corresponding to Africa, Eurasia (Europe, Middle East, and Central/South Asia), East Asia, Oceania, and the Americas—an intercluster population pair is plotted only if it includes one population from Africa and one from Eurasia, one from Eurasia and one from East Asia, or one from East Asia and one from Oceania or the Americas.

For population pairs from the same cluster, as geographic distance increases, genetic distance increases in a linear manner, consistent with a clinal population structure. However, for pairs from different clusters, genetic distance is generally larger than that between intracluster pairs that have the same geographic distance. For example, genetic distances for population pairs with one population in Eurasia and the other in East Asia are greater than those for pairs at equivalent geographic distance within Eurasia or within East Asia. Loosely speaking, it is these small discontinuous jumps in genetic distance—across oceans, the Himalayas, and the Sahara—that provide the basis for the ability of STRUCTURE to identify clusters that correspond to geographic regions.”

Leszek said: If you look at geographical distribution of a genetic traits using a map, you often will notice that the distribution doesn’t really line up with the boarders of races.

Look up Steve Hsu’s post on 40 years of clustering (also see his posts on European genetic substructure, and scientific basis for race).

“Represent each individual human by their DNA sequence. When aggregated, they cluster into readily identifiable groups. This has been known for 40 years now, although the technology and methods of analysis continue to improve. Below are results from 1966, 1978 and 2008.”

http://infoproc.blogspot.co.nz/2009[…]rogress.html

D P Robin said:

Any of us, who took Biological Anthropology at the University of Michigan, from the 60’s until a year or so ago, had the chance to take C. Loring Brace’s course on race, which made most, if not all these points. The faculty in my time there (mid-late 1970’s) had little good to say about the concept of human races at all. It was hard to cling to ideas that races were “real” after hearing Conrad Kottak talk about racial taxonomy in Brazil (where he had done fieldwork). dpr

Anthropologist recalled this conversation with Loring Brace:

“One commenter asked whether Loring Brace actually believes what he says and writes. I’m sure he does. ‘Belief’ is rarely the result of careful thinking and impartial weighing of pros and cons. People want to feel good about themselves, and often they prefer feeling good and being wrong to feeling bad and being right.

I remember corresponding with Loring Brace about Lewontin’s finding that genetic variation within human populations greatly exceeds genetic variation between human populations. I pointed out that we see this same genetic overlap between dog breeds that are nonetheless phenotypically distinct. He replied that dog breeds are a creation of human-directed selection and, thus, irrelevant.

I then pointed out that not all dog breeds have been created by kennel clubs. More to the point, there are many sibling species that show the same kind of genetic overlap and yet are distinct in anatomy and behavior.

At that point, he backed off completely. He said that genetic arguments were not critical to this issue anyway. I also had the creepy feeling that he wasn’t really surprised by what I said.”

http://westhunt.wordpress.com/2011/[…]ck-russians/

With regard to acoolcoolhippo, would like to throw a couple of recent studies into the mix. On the distinction between Sub-Saharan Africans and non-Sub-Saharan Africans, the 2009 study by Long et al. Human DNA sequences: More variation and less race finds non-Sub-Saharan Africans to be a nested subset of the larger diversity already contained within Sub-Saharan Africa, and that this Sub-Saharan African genetic diversity could in some ways also cluster distinctly: “The pattern of DNA diversity is one of nested subsets, such that the diversity in non-Sub-Saharan African populations is essentially a subset of the diversity found in Sub-Saharan African populations. The actual pattern of DNA diversity creates some unsettling problems for using race as meaningful genetic categories. For example, the pattern of DNA diversity implies that some populations belong to more than one race (e.g., Europeans), whereas other populations do not belong to any race at all (e.g., Sub-Saharan Africans).”

On craniometrics I respect the careful work of Relethford cited above. However, a more recent assessment, Strauss and Hubbe 2010, Craniometric Similarities Within and Between Human Populations in Comparison with Neutral Genetic Data suggests that the craniometric clusters may not be quite as evident as they have been portrayed: “Contrary to what was observed for the genetic data, our results show that cranial morphology asymptotically approaches a mean ω of 0.3 and therefore supports the initial statement–that is, that individuals from the same geographic region do not form clear and discrete clusters–further questioning the idea of the existence of discrete biological clusters in the human species.”

Many participants in this exchange, and articles like Templeton’s that people are using to support their ideas, appear to think that Fst or Gst are measures of genetic divergence. There are even discussions of cut-off values of Fst = 0.25 to 0.75 to define subspecies. This is crazy, since the maximum possible value of Fst for a given locus is constrained by the mean within-group heterozygosity at that locus. If within-group variation is high, it is mathematically impossible for Fst to reach those thresholds (or any given threshold), even if the groups cannot interbreed, share no alleles, and are each on their own evolutionary course. See Hedrick 2005 or Jost 2008 for the math. Fst and its relatives can approach zero even if no alleles are shared between groups, and can equal unity even if almost all groups are fixed for the same allele.

Also, some commentators on other threads reacting to Coyne’s post have claimed that subpopulations would only diverge genetically at neutral or nearly neutral alleles if the absolute number of migrants per generation were less than 1, based on standard population genetic arguments. Since such a low value of migration is improbable in humans, they argued that human subpopulations would not diverge at neutral or nearly neutral loci. This argument is ultimately based on linking the number of migrants to the equilibrium value of Gst or Fst under simple genetic models. It rests on the misinterpretation of Fst as a measure of genetic differentiation ranging from 0 to 1. As I just mentioned, Fst is not a measure of differentiation in this sense. Therefore the standard population genetic argument used by these commentators is wrong. As I show in Jost 2008, the real quantity determining whether two subpopulations coheres or diverges genetically is m/u where m is the relative migration rate and u is the mutation rate of the locus in question. Divergence does not depend on the absolute number of migrants but on the relative migration rate; (number of migrants) / (group population). Thus oceans, mountains, deserts, and other leaky barriers might have been sufficient to allow divergence of subpopulations, especially when the subpopulations are large and migration channels are narrow.

I would urge all parties to try to avoid using Fst, and to be cautious about using pop gen arguments that depend on Fst, in this discussion.

Hedrick P (2005) A standardized genetic differentiation measure. Evolution, 59, 1633-1638. Jost L (2008) Gst and its relatives do not measure differentiation. Molecular Ecology, 17, 4015-4026.

Jason Antrosio said:

With regard to acoolcoolhippo, would like to throw a couple of recent studies into the mix. On the distinction between Sub-Saharan Africans and non-Sub-Saharan Africans, the 2009 study by Long et al. Human DNA sequences: More variation and less race finds non-Sub-Saharan Africans to be a nested subset of the larger diversity already contained within Sub-Saharan Africa, and that this Sub-Saharan African genetic diversity could in some ways also cluster distinctly: “The pattern of DNA diversity is one of nested subsets, such that the diversity in non-Sub-Saharan African populations is essentially a subset of the diversity found in Sub-Saharan African populations. The actual pattern of DNA diversity creates some unsettling problems for using race as meaningful genetic categories. For example, the pattern of DNA diversity implies that some populations belong to more than one race (e.g., Europeans), whereas other populations do not belong to any race at all (e.g., Sub-Saharan Africans).”

On craniometrics I respect the careful work of Relethford cited above. However, a more recent assessment, Strauss and Hubbe 2010, Craniometric Similarities Within and Between Human Populations in Comparison with Neutral Genetic Data suggests that the craniometric clusters may not be quite as evident as they have been portrayed: “Contrary to what was observed for the genetic data, our results show that cranial morphology asymptotically approaches a mean ω of 0.3 and therefore supports the initial statement–that is, that individuals from the same geographic region do not form clear and discrete clusters–further questioning the idea of the existence of discrete biological clusters in the human species.”

@ Jason the hippo character has asked me to pass on these comments:

You bring up a great example of why it’s important to specify a subspecies concept — along with this concept’s criteria. How does Long et al.’s critique relate to the classification of human populations by the geographic race concept being discussed? (Refer to my discussion above.) Maybe Long et al.’s point is relevant when it comes to some other subspecies concepts, but it is not when it comes to this one.

However, a more recent assessment, Strauss and Hubbe 2010, Craniometric Similarities Within and Between Human Populations in Comparison with Neutral Genetic Data suggests that the craniometric clusters may not be quite as evident as they have been portrayed: “Contrary to what was observed for the genetic data, our results show that cranial morphology asymptotically approaches a mean ω of 0.3 and therefore supports the initial statement–that is, that individuals from the same geographic region do not form clear and discrete clusters–further questioning the idea of the existence of discrete biological clusters in the human species.”

If you noticed, above I said:

But what about the 75% differing interpretation? This is where I get caught up. Craniometric and dental variation is such that while you can use it correctly classify individual, the differences aren’t so extreme that 75% of one population differ from 97% of another even using any combinations of the traits; it’s not even close. These then are not good differentiating criteria.

The reason I said this is because I was aware of Strauss and Hubbe’s study. The findings seem to imply that in no set of craniometric differences are more than 75% of one population different from 97% of another. It would be interesting to run the same analysis using the dental data or skeletal data or a combination of all. Whatever the case, the situation illustrates the important difference between two main interpretations of the 75% rule. One can correctly classify more than 75% of individuals based on a set of differentia, without 75% of individuals lying outside 97% of the range of the others in the same differentia.

Anyways, this is why I offered natural hair curliness as a differentia (for Subsaharan Africans and non-Africans.) Or was that too superficial?”

@ Nick & Jason,

Chuck has asked me to add one final comment as follows:

“Last comment.

Nick and Jason,

I would agree that human populations are at best only weakly subspeciated. As such, if you tighten up the taxonomic standards, likely, no set will qualify as taxonomic races. But, as it is, the standards for non human populations are fairly lax. Let me quote from Remsen (2010), “Subspecies as a meaningful taxonomic rank in avian classifications”: Mayr et al. (1953) provided objective, quantitative definitions of subspecies based on degree of overlap that can be applied across taxa. They outlined why using simple linear overlap in measurements, for example, overemphasizes extreme individuals in a population and overestimates true population overlap. They also discussed various interpretations of the “75% rule” as the threshold for naming subspecies. Although one interpretation is that only 75% of the individuals of each sample have to be correctly classified, the rule as defined by Amadon (1949), Mayr et al. (1953), and Patten and Unitt (2002) is based on standard deviations from the mean of normally distributed data. Depending on which metric is applied, in essence these definitions mean that 90–97% of the individuals of one population must be distinguishable from the equivalent percentage of the other population to be considered subspecies under the somewhat misleadingly named 75% rule.

As I was noting above, there are multiple interpretations of the 75% rule. The more lax interpretation is (or was) just that 75% of individuals can be correctly classified into the respective populations. This is the interpretation that Sewall Wright (quoted above) was referring to in “Evolution and the Genetics of Populations” in his discussion of human races. Ditto Bodmer and Cavalli-sforza in “Genetics, Evolution and Man.” By this reading of the rule, clearly the 5-7 major human population clusters qualify as subspecies. I can’t imagine anyone seriously dissenting on this point. Perhaps, they would argue that by event of mass transportation many individual don’t “share a unique geographical range ” with others of their said race and maintain that this invalidates the concept as applied to humans. This is an interesting point — or would be were someone to make it, but this would, at best, only regress the question from “Are the said populations subspecies?” to “Were the said populations recently subspecies?”

Continuing…. […]Although the 75% rule has a long history in ornithology, its application has been erratic at best. For example, it is generally not mentioned as a criterion for recognizing subspecies in classifica- tions (e.g., American Ornithologists’ Union 1957, Dickinson 2003) or in any of the Handbook of the World series (del hoyo et al. 1992–2008). It is not possible to tell how many of the subspecies currently recognized in such sources would qualify as subspecies under the 75% rule, but it is certain that many subspecies, especially in North America, would not qualify as valid taxa under this rule, particularly those defined by mensural differences. From personal experience in attempting to use subspecies diagnoses, such as the keys in the Birds of North and Middle America series (ridg way and Friedmann 1901–1950), I predict that more than 75% of North American subspecies taxa delimited by mensural data would not survive application of the 75% rule

Now, by this more rigorous interpretation of the rule, it’s not clear if the 5-7 major human population clusters qualify using any combination of traits — at least, I’m not aware of the qualifying differentia. They might be out there, though. I suggested hair curliness as a differentia between non-Sub Saharans and Sub Saharans. Based on the means and standard deviations presented in Hrdy’s “Quantitative Hair Form Variation in Seven Populations,” for “hair curvature,” these two populations seem to meet the 75% (differentia) criteria. The among to within F-ratio for this trait was 365. And the results agree with that of Loussouarn (2007), assuming curliness and curvature refer to basically the same trait. I’m sure if pressed, though, I could find other traits.

Whatever the case, going by Remsen (2010), it doesn’t seem as if the 75% (differentia) rule is rigorously applied. Why should it be de rigueur, then, when it comes to humans? Can one really maintain that describing human populations as subspecies is an affront to the subspecies concept, give how the concept is actually employed?

About this Entry

This page contains a single entry by Nick Matzke published on February 29, 2012 2:38 PM.

Update on Springer “Biological Information: New Perspectives” Volume was the previous entry in this blog.

Freshwater: RNCSE article published online is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Categories

Archives

Author Archives

Powered by Movable Type 4.38

Site Meter