ps418

This is a critique of an article written by Tina Hesman, dated 5/13/03, titled DNA evidence shows race doesn't exist. The central conclusion of the article is that "[t]he recently completed Human Genome Sequencing Project has confirmed what many scientists knew all along — that humans don't fit the biological criteria that defines race." The article makes a number if claims which, regardless of what one thinks of the usefulness of the concept of race, are simply inaccurate and misleading. I argue that the article invokes a faulty and and idiosyncratic definition of race, incorrectly defines crucial genetic terms, and misprepresents the evidence relating to human genetic differentiation. I argue further that, contrary to Hesman's assertion that scientists agree that there are no human races, on the most widely-used definitions of race or subspecies, human races/subspecies do objectively exist.

To begin, there are a few points of agreement. First, social definitions of race based on simple skin color are often inaccurate and do not correlate well with genetic data, so in that specific sense genetic data has indeed undermined the concept of race -yet obviously one can dismiss folk race taxonomies yet still accept that biologically meaningful human races exist. Second, there is so much variation within any population, particularly the continent-scale groups called "races," that even to the extent that race is a valid concept, it is in most cases of very limited utility in predicting, say, an individual's genotype or phenotype. Not necessarily zero utility (e.g. in absence of individual geneotype, may be useful as a predictor of drug response; Risch 2002), but definitely very limited utility. Third, there has always been some level of gene flow between populations, so there are no "pure" races or races as incipient species: many causcasoids possess african genetic admixture, and vice versa, for instance. With the rise of the modern economy, the natural impediments to gene flow have become even less relevant. Fourth, whether there are human races is like everything else definition-dependent, and one is free to embrace any definition one wishes. Having said this, there is much in Hesman's article to disagree with. For instance, Hesman states:

First, the "some scientists" appears to refer solely to Alan Templeton, who is cited in Hesman's article. Few if any other researchers require any particular Fst value for the diagnosis of subspecies (Goodrum, 2002; race and subspecies are hereafter used as synonyms). There are many definitions of subspecies in use, most emphasizing geographic distribution and morphological distinguishability, and none of them as far as I know require a particular level of genetic divergence. If they did, then subspecies could not be recognized at all in the absence of genetic data. Second, the definition of Fisher's Fst given by Hesman is simply incorrect -- Fst is a measure of population subdivision, not simply of heterozygosity. And third, there are in fact many widely recognized subspecies with autosomal Fst values much lower than 0.25.

So, the definition of race offered by Hesman seems highly idiosyncratic. In the words of Kaplan and Pigliucci (2002), "many of the current arguments against the existence of biologically meaningful human races fail precisely because they rely on a use of the biological race concept that is not in fact in wide circulation." But if Fst doesn't define race/subspecies, then what does? There are a variety of definitions. For instance:

Hartl and Clarke define race/subspecies as “a group of individuals in a species showing closer genetic relationships within the group than to members of other such groups” (p. 301).

Mayr (1996) defines race/subspecies as "conspecific populations that differ from each other morphologically.”

The Oxford Paperback Reference Concise Science Dictionary (1996) defines race/subspecies as a "group of individuals within a species that breed more freely among themselves than with other member of the species and resemble each other in more characteristics."

King and Stansfield (1990) defines race as a "phenotypically and/or geographically distinctive subspecific group, composed of individuals inhabiting a defined geographical and/or ecological region, and possessing characteristic phenotypic and gene frequencies that distinguish it from other such groups. The number of racial groups that one wishes to recognize within a species is usually arbitrary but suitable for the purposes under investigation.”

Using another, closely-related definition of subspecies --what is known as the 75% rule-- also results in the recognition of human subspecies. The rule says that "75 per cent of the individuals classified in one subspecies are distinguishable from 100 per cent of the individuals belonging to the other subspecies of the same species, which is statistically equivalent to 90 per cent joint non-overlap” (Groves 1989, p. 7). For instance, using only a few genetic markers allows greater than 75% accuracy in determing an individual's continent of origin (e.g. Africa, Asia, Europe), while the use of greater numbers of markers results in ~99% accuracy (e.g. Shriver et al, 1997; Rosenberg et al, 2002; Bamshad et al, 2003). The same is also true for skeletal morphology. Individuals can be assigned to continental groups with much better than 75% accuracy based on craniofacial morphology (Gill, 1998; see also Hanihara et al, 2003). By these definitions, which are in far wider use than the definition mentioned by Hesman, human subspecies objectively exist. In the words of Ernst Mayr (2002, p. 90):

Hesman continues:

Here we come to the most serious error in Hesman's article. The specific Fst value depends on what set of genetic data is used to compute it, for instance mtDNA, Y chromosome DNA, or autosomal DNA. Comparisons are only meaningful when they are comparisons of similar data. However, the Grey Wolf Fst value of 0.7 is based on mitochondrial data, whereas the human Fst is based on autosomal data. Why is this a serious error? This is a very serious error because the lower effective population size of mtDNA and Y chromosomes ensures that differentiation due to drift will be much greater for these systems than for autosmal DNA (the vast majority of genetic material in animals is of course autosomal). And in fact, the autosomal Fst value for Grey Wolves is only 0.168 (Roy et al, 1994), which in fact is nearly identical to human autsomal Fst value of ~0.15, and several (~5) Grey Wolf subspecies are recognized (Grey Wolf). Furthermore, human global Fst values computed from mtDNA (0.30-0.35) and Y-chromosome genetic data (0.35-0.60) are well above 0.25 (Jorde et a, 2000), so using these genetic systems does result in the recognition of human races according to Hesman's criterion. Graves himself, who is cited in the article, has said that "Gray wolves are strongly subdivided" (source), so it is difficult to see how he could deny that human populations are at least moderately subdivided, much less claim that there is essentially zero genetic differentiation of human populations (see below).

Hesman writes:

Again, Fst is not simply a measure of diversity, and an Fst score of 0.08 does not mean that 8 percent of genes have more than one allele. Ironically, immediately after stating that global human Fst is 0.15, we're now told that some disparate populations have Fst scores of 0.25! This should have alerted the author that apples (autosomes) and oranges (mtDNA) are being compared. And though the article admits that such a score qualifies the two populations for subspecies designation, it is then said that these poplations cannot qualify because they are not geographically isolated. However, geographic isolation is neither a valid nor a useful criterion for "subspecies". It would be hard to avoid the conclusion here that 'race' is simply being defined out of existence

And the claim that average between-population Fst is 0.08 is highly misleading, as this value depends on which genetic data is used and which populations are compared. For instance, a comparison of Chinese and Tibetan populations will probably yield low Fst, but then no one has ever suggested that these two populations represent different human subspecies or races! When you compare Fst between continental populations, which more or less is where race proponents would draw the subspecies divisions, the values are much higher than 0.08 (e.g. Jorde et al, 2000). Further, a comparison of several sets of genetic data by Jorde et al shows that "the differentiation of individual populations within continents is several times lower than the differentiation between continental populations," and thus continent-level human population groups clearly fulfill Hartl and Clarke's genetic definition of subspecies given above.

Hesman's article cites Joseph Graves:

That statement is factually incorrect. The truth is that human populations show a moderate degree of divergence compared to other mammals. We are neither exceptionally subdivided nor exceptionally homogenous. To state that the divergence between people is essentially zero is a direct contradiction of the available evidence, not an inference from it. In fact, even assuming a global Fst value of 0.15, according to Fisher himself, this indicates moderate/great genetic differentiation, while only a value of 0 to 0.05 would indicate little differentiation.

No matter what type of genetic data you use, Fst of human populations within and between continents are consistently and signficantly greater than 0 (e.g. Jorde et al, 2000). So, Graves has it precisely wrong: no matter how you measure it, the amount of divergence between populations traditionally assigned subspecies status is never "essentially zero."



Refs

Bamshad et al., 2003. Human population genetic structure and inference of group membership. American Journal of Human Genetics 72, 578-590.

Gill, G.W., 1998. Craniofacial criteria in the skeletal attribution of race. In Forensic Osteology: Advances in the Identification of Human Remains, ed. by Kathleen J. Reichs, 293-317. Springfield: Charles C. Thomas Publisher, Ltd.

Goodrum, J., 2002. The Race FAQ. Accessed 5/21/03 at: http://www.goodrumj.com/RaceFaq.html

Groves, C.P., 1989. The biology of race. Berkeley: University of California Press.

Hanihara et al, 2003. Characterization of biological diversity through analysis of discrete cranial traits. American Journal of Physical Anthropology. Published Online: 13 Mar 2003

Hartl, D., Clark, A.G., 1989. Principles of population genetics. Sinauer Associates, Sunderland, MA.

Kaplan, J., and Pigliucci, M., 2002. On the concept of biological race and its applicability to humans. Accessed 5/21/03: http://philsci-archive.pitt.edu/archive/00001078/

King, R.C., and Stansfield, W.D., 1990. A dictionary of genetics. Oxford University Press. 4th ed.

Mayr, E., 1996. What is a species and what is not? Philosophy Of Science 63, 262-277.

Mayr, E., 2002. The biology of race and the concept of equality. Daedalus, Winter 2002, 89-94.

Risch, N., 2002. Categorization of humans in biomedical research: genes, race and disease. Genome Biology 3(7):comment2007.1-2007.12 .

Rosenberg et al, 2002. Genetic structure of human populations. Science 298, 2381 - 2385

Roy et al, 1994. Patterns of differentiation and hybridization in North American wolflike canids, Revealed by Analysis of Microsatellite Loci. Molecular Biology and Evolution 11, 553-570.

Shriver et al, 1997. Ethnic-affiliation estimation by use of population-specific DNA markers. American Journal of Human Genetics 60, 957-964.

Friar Bellows

Interesting article. Here's the conclusion from the FAQ you referenced: