The Genetics of Race

Everyone has heard the fashionable view that race is not a valid biological concept but is, instead, a suspect sociological category. The implication is that racial distinctions are a form of superstition, that there is no justification for wanting to preserve races, and that replacement of one group by another represents no genetic loss. More specifically, it implies there is no reason for whites in Europe or America to resist displacement because we are, in effect, being replaced by ourselves.

This view is wrong on its face. Races breed true, generation after generation — Danes cannot produce a Bantu, and Bantus cannot produce a Dane — and racial differences are so substantial and consistent that two-year-old children notice them. Scientists recognize the biological differences between animal subspecies — eastern lowland and mountain gorillas, for examples — that are far more physically similar to each other than are members of different human races.

Clearly, the main motive for promoting a view so contrary to common sense is “anti-racism,” but there is one biological fact on which the race-deniers build their argument: This is the oft-cited observation by Richard C. Lewontin of Harvard that there is far more genetic variation within human racial groups (about 85 percent of the total) than between them (about 15 percent). This is true, and scientists were surprised when this fact first came to light. There have since been many outrageous misinterpretations of Prof. Lewontin’s findings, with some people even claiming that because there is more genetic variation within than between human groups, whites are genetically more similar to blacks than to other whites.

I will try to explain what the Lewontin findings mean, and how they should be understood. To do so, I will simplify some of the basic concepts of population genetics and use analogies to illustrate certain key points. Readers interested in the more technical aspects of the subject can consult the papers mentioned in this article, and an Internet search will uncover a vast and challenging literature. What follows, however, should be enough to clear the cobwebs.

First, what is “genetic variation?” Genes are made from DNA, which is made of four chemical elements called nucleotides (Adenine, Cytosine, Guanine, and Thymine, abbreviated as A, C, G, T). These nucleotides are arranged in ordered combinations (e.g., ACGTCGATGATGCA) that make up DNA sequences. DNA is the code in which cells store the information about how to build the basic chemical components of the human (or any other) body. Through a complex chain of events, the information in these sequences is first “transcribed” into an intermediate form called RNA, and then “translated” into proteins, which are the building blocks of life. Some of the DNA sequences regulate transcription, and what subsequently happens to the RNA. Specific DNA sequences with specific functions are called genes, and the complete set of DNA sequences of an organism is called its genome.

A large portion of the human genome is said to be “non-functional,” in that the DNA sequences do not code for proteins nor do they seem to have any regulatory functions. Recent studies suggest that some sequences previously thought to be non-functional may be involved in regulation, but scientists are interested even in DNA sequences that are truly neutral and non-functional. This is because these sequences sometimes vary in interesting ways between individuals and groups, and population geneticists study them to learn about ancestral relationships. Thus, when I refer to genes and genetic variation, I am referring to all types of DNA sequences that can be the source of genetic variation. How do the sequences vary?

There are differences between individuals and between races and ethnic groups. Sometimes these differences have profound effects on function and physical form (phenotype) and sometimes they do not. Individuals differ in many ways — height, coloring, intelligence, personality — because of differences in their DNA sequences, though environment also affects these traits to varying degrees. Different information stored in the DNA produces these differences. In other words, variants of the same genes can produce different effects in different people.

There are further differences between people in the so-called neutral (non-functional) gene sequences, but these do not have physically apparent or visible effects. Siblings may be very similar to each other in all apparent ways, and are similar in those portions of DNA that code for detectable differences. At the same time, there may be considerable differences in their non-functional DNA, and these differences do not show up as apparent, or phenotypic differences. Genetic variations can have no effect or very profound effects, depending on which genes they affect. In some crucial sections of the genome, even the slightest variation can be very damaging. A few conditions, such as cystic fibrosis, sickle cell anemia and alpha-1-antitrypsin deficiency, result from tiny differences in gene sequences at certain key points.

There are a few genetic variations that are population-specific, which is to say that they are found in one population but are essentially absent in another (see the note at the end of this article), but most differences are in frequency: One variation of a gene is found more or less frequently in one group than in another. In some cases there are many variants of a gene sequence, in some cases, only two variants. For example, a single gene may have three variants. In population A, gene variants 1, 2, and 3 may have frequencies of 30 percent, 30 percent, and 40 percent (for a total of 100 percent), while in population B the same frequencies may be 20 percent, 60 percent, and 20 percent and in population C, 70 percent, 25 percent, and 5 percent.

People are often surprised to learn that most of the human genome does not have racial or ethnic patterns. Most genetic variation is “random,” which is to say that members of different population groups are equally likely to have different variants of most genes. This is particularly common in those parts of the genetic structure that appear to be non-functional, and do not seem to have an effect on the organism. The sum of all these genetic differences within the human species is the total genetic variation of humans.

This total genetic variation can therefore be divided into the differences in gene frequencies found between individuals from the same group, and the differences in gene frequencies found in different groups. The portion of the genetic variation in which there are random differences between individuals of the same race is larger than the portion that is patterned by race. As we saw above, the proportions are approximately 85 percent and 15 percent, which is to say that 85 percent of the genetic variation among people of the same race is equally random when compared to people of different races. For all this variation, therefore, there are no patterns that indicate whether someone is a Pygmy or an Eskimo or a European.

It is in this sense that Prof. Lewontin is correct: Within each racial group, there are more DNA sequence variants that are random than there are DNA sequence variants that show a racial pattern. It is therefore correct to say there is more genetic variation within races than between them. The anti-racists twist this fact to imply that individuals of different races are (or can be) more similar to people of other races than to people of their own race.

In fact, since there are no racial patterns to 85 percent of human genetic variation, that is theoretically possible. Purely random variation in these areas could conceivably make two individuals of different races more alike than two individuals of the same race. However, in the remaining 15 percent — the genetic variation where consistent racial differences are found — they would be as different from each other as any two typical members of the different races. Theoretically, a Chinese could be found who was indistinguishable from a Frenchmen in large parts of their DNA, but this would not make them particularly similar. This Chinese would not have the gene variants that contribute to producing light-colored eyes or hair, or Caucasian facial features, for example.

The following crude analogy does not capture the complex, real world of human genetic variation, but it does demonstrate the importance of small sets of structured data compared to a large body of random variation. Let us imagine a group of 100 boys and 100 girls, each with 100 marbles that can be any color. For the first 85 marbles, color differences are completely random for everyone. However, for the last 15 marbles, boys always have blue marbles and girls have pink. Boys and girls thus form two distinct groups based on 15 percent of the marbles; the other 85 percent of the marbles vary in color without regard to sex. This means there is more intra-group variability than inter-group variability, but this does not mean that the blue-for-boys and pink-for-girls pattern does not exist; it clearly does. Nor does it imply that any given boy is likely to have marbles that are more similar to those of a girl than to another boy. The opposite is likely to be true.

On the other hand, as in the case of the Chinese and the Frenchman, it is theoretically possible, by random chance, that a specific boy’s first 85 marbles will be so similar to a specific girl’s that this boy will have a total set of marbles more similar in color to that girl’s than to any other boy’s. Theoretically, a given boy and girl could have 85 marbles that were exactly the same color. Does this mean the two larger groups do not exist? No. There are still two groups clearly defined by the consistent differences in color of the last 15 marbles, while the first 85 are random. But if marble color is a criterion for putting children into groups, shouldn’t the boy and girl be classified together as members of a group? No, because their similarity is a result of pure chance. What is significant, and what makes the two groups meaningful is the pattern of difference in just 15 percent of the marbles.

There are several ways to look at genetic variation within and between races. All humans, for example, share a tremendous number of traits. We all have two arms and legs, one heart with four chambers, and one stomach. We all perform complex DNA replication and protein translation, and we process sensory information in the brain. In this sense, the differences between races are tiny. Much of the genome is taken up with information on how to build the parts of the body common to all people, and a lot of the rest is inactive. In the case of genes that build the parts and processes common to all humans, there are random variations that do not seem to make any real difference in, say, how a liver or stomach works. That part of the genome varies more or less randomly across the entire human population, as do the parts that seem to do nothing. It is in the relatively small part of the genome that produces the relatively small differences that distinguish races that we find the consistent genetic patterns that make Tibetans different from Maoris. Given the vast biological ways in which Tibetans and Maoris are similar, it does not take much of the genome to produce the relatively small ways in which they are different.

The same is true for the distinctions between humans and other species. Some 98 percent of the chimpanzee genome is indistinguishable from the human genome, and humans share no less than 80 percent of their genes with mice. How does the 98 percent figure square with the 85 percent/15 percent figures for within-group and between-group genetic variation? However much humans vary genetically from each other, they all vary in ways specific to our species; the two percent difference between humans and chimps represent the small areas where there is no overlap at all between the two genomes, and that is enough to account for the differences between the two species.

A different way to compare variation within a race to variation between races is to consider the African American/European American IQ distribution curves. The two groups differ by one standard deviation, or 15 points. Within each group, the variability in IQ is much greater than 15 points — from moron to genius — but that does not mean a random white is more likely to have the same IQ as a random black than a random white. The opposite is true. Nor does the great variability in IQ within races somehow diminish the importance of the much smaller, 15-point difference in average black and white IQs. Of course, there is significant overlap, so that for this one trait, some whites and some blacks are indeed more similar to each other than they are to members of their own racial group.

The similarities drop away when we compare the two races on a larger number of traits. Imagine that for each trait, there is more variation within the group than between groups; this may well be true given the relatively large range of admixture and phenotypes found in African Americans and, to a lesser extent, in European Americans. For each of these traits taken individually — complexion, certain aspects of facial features, behavior, body form, skull shape, metabolism, etc. — there will be some degree of overlap between the groups, but the level of variation within the groups will be greater than the average variation between groups. (This is a comparison of American blacks and whites, between whom there has been considerable mixture. A comparison between black Africans and Europeans would find no overlap in some traits, such as complexion or facial features.) However, when all the traits are taken together, the chances of overlap between the two groups approach zero; one will not find whites and blacks who overlap on all, or even most, of a large number of phenotypic traits. For a single trait, or a small number of traits, we may see overlap, but not for 20 or 50 or 100 traits taken together.

The same principle applies to genetic data. Looking at genes in isolation leads to Prof. Lewontin’s conclusion, but as soon as one looks at numbers of genes, distinct groups emerge and are easily classified (see next article, page 7). Furthermore, functional genes with evolutionary importance may be concentrated in the parts of the genome that are unique to different populations. When sharp environmental differences began to push groups in different evolutionary directions, the resulting changes in the genome were very small, but had great phenotypic significance. These would include genes that account for differences in physical appearance, as well as differences in other genetic variants that may have physiological significance. Race differences in disease rates and medicinal effect also reflect evolved genetic differences between the races. The large amount of genetic variation that does not follow a racial pattern should not blind us to the small but very important part that does.

Readers with a technical bent can read the online article by A.W.F. Edwards called “Human genetic diversity: Lewontin’s fallacy” (BioEssays 25:795-801, 2003), in which Prof. Edwards mathematically dissects and refutes the “race does not exist” fallacy. Prof. Edwards explains that Prof. Lewontin’s conclusions are “unwarranted because most of the information that distinguishes populations is hidden in the correlation structure of the data and not simply in the variation of individual factors.” What this means is that looking at individual genes is not enough to distinguish between populations; groups of genes must be taken together.

Here is an example from the Edwards paper: Imagine two populations in which the frequency of a particular gene variant compared to its alternative is 70 percent in population A and 30 percent in population B. If you use this one gene as the criterion for determining who is an A, you will mistakenly classify 30 percent of Bs as As because 30 percent of them have the variant that is more common in As. As you increase the number of variants you are comparing, the chances of misclassification decrease because it is increasingly unlikely for a member of one group to have a distribution of many different gene variants similar to that of the other group.

(How do geneticists determine how much variation is within groups or between groups? In the case of the example in the text, with 70 percent of population A and 30 percent of population B having one of two gene variants, within-group variation as a percentage of total variation is calculated as follows: 4 × 70 percent × 30 percent = 84 percent. Although a 70:30 and a 30:70 split in gene variations seems like a big difference between groups, only 16 percent (100-84 = 16) of the genetic variation is considered to be between rather than within groups. In that specific respect, the two groups are more similar than they are different.

What if the split is 90:10 and 10:90? The calculation for within-group variation is 4 × 90 percent × 10 percent = 36 percent. Even with a split as substantial as nine to one, only 64 percent of the total variation is between-group. If the variation is 99:1 and 1:99, the within-group variation is still four percent, though the between-group variation rises to 96 percent of the total variation.)

We find the same principle in opinion polling. Statistics show that the larger the sample size, the more accurate the poll. If you want to know what percentage of Republicans support the Iraq war, you will get more accurate results by interviewing 1,000 registered Republican than by interviewing three, especially if one of them is Pat Buchanan. The same holds true for genetic classification. One must “poll” a large number of gene variants to distinguish population groups, because there is so much genetic variation within the groups.

This large amount of intra-racial variability does not at all imply that whites are more similar to blacks than to other whites. It means simply that of the total amount of human genetic variation, most is randomly distributed among individuals, with no racial trends. In these areas whites are, in general, neither more similar to whites than to blacks, or to blacks than to whites; the variation is random for everyone, with no racial pattern. There is a fraction of genetic variation, however, that is highly structured at the population level, meaning whites are much more similar to whites than to blacks (or any other race) for this important part of the variation.

One more analogy may help demonstrate the great importance of looking at genetic distinctiveness rather than at overall genetic similarity. A person’s gametes (sperm and eggs, which have only half, or one “copy” of a person’s genome) are genetically different from each another. By chance, one of Joe’s sperm could have a greater overall genetic similarity, including non-functional DNA, to one of Ted’s sperm than to another of Joe’s own sperm. Does this mean Joe is more genetically similar to Ted than to himself? Does it mean Joe should be indifferent to whether Ted’s sperm are used to create Joe’s child? Of course not. In comparing Joe and Ted, the important genetic information is in the parts of the DNA that distinguish Joe from Ted, not in the random genetic variation — much of it which has no functional significance — that is found in all individuals.

The same is true when comparing different population groups. If a species of wolf were threatened with extinction, environmentalists would not dismiss the problem by saying “there is more genetic variation within each wolf species than between them,” or “the threatened species shares most of its genetic variation with other wolf species and even with dogs, so there is nothing to worry about.” Extinction of any group, including human races, means the irreplaceable loss of unique characteristics and unique genetic information that distinguish that group from all others.

Let us consider further the within-group genetic variation that Prof. Lewontin and his supporters make so much of. This “within-group” variation is present in all human populations. If all humans except Bushmen became extinct, then the bulk of human genetic variation would survive in those Bushmen. Is then nothing lost? Hardly. All of the genetic variation that distinguishes the other peoples of the world from both Bushmen and from each other would be lost. It is true that the lost fraction of variation would be considerably smaller than the fraction that remained, but does that mean the lost fraction is unimportant? No. It is this much smaller fraction of human variation that makes different groups unique. It is therefore inherently much more valuable. The same applies to any endangered species; the potential loss is not what the species has in common with other species, but what makes it unlike all the others.

Natural selection works on differences between organisms, not similarities. If all members of a species were identical clones, differences in survival rates in a given environment would be a matter of pure chance. However, genetic differences result in physical differences, which result in different outcomes and different survival rates in different environments. As Vince Sarich and Frank. Miele note in their book Race: “Simply stated, the case for race hinges on recognition of the fact that genetic variation in traits that affect performance and ultimately survival is the fuel on which the evolutionary process runs.” In other words, racial differences are real and significant: They help different groups survive in different environments. It is the unique, patterned genetic differences of groups that affect whether their members are more or less likely to survive; not the great bulk of their gene sequences, which they hold in common with all other members of the species.

Some followers of Prof. Lewontin try to argue that races cannot exist because there is variability between the sub-races within races. Thus, they argue there is no such thing as a black African race because there are substantial differences between West Africans and East Africans. This argument misses the point. There are many different populations one could call races, depending on how the term is defined. It would not be incorrect to call West Africans and East Africans different races. Someone could call the French and the Poles different “races,” so long as he was consistent in calling other groups “races.” However, at the global level, Africans cluster with Africans, and Europeans with Europeans; they constitute legitimate racial groups.

Vdare.com correspondent Steve Sailer has pointed out that the same logic applies to the concept of location. For example, Tampa is not the same place as Miami. However, compared to Cleveland and Cincinnati — also two separate locations — Tampa and Miami cluster together in Florida. Cleveland and Cincinnati are in Ohio. Both Florida and Ohio can be said to cluster in the United States. When it comes to human races, their number varies greatly depending on how they are defined. Most people define races at the continental level, and define different groups within the same continental race as “subraces” or “ethnic groups.”

Some people argue that all races merge gradually into all others at their borders, and that this means race does not exist. First, this is not true. Gradual shifts from one race to another can be found within a continent, but between continents there are distinct racial breaks. This helps substantiate the usual definition of race at the level of continents. At the same time, the existence of mixed-race people does not, somehow, negate the existence of races. It substantiates it. There could not be mixed-race people if race did not exist.

Race is therefore a concrete, objectively-determined biological fact. Race is a “social construct” only insofar as a society may label people using criteria that are at least partially independent of biological reality. For example, in America, “blacks” include people with a wide range of ancestry, and some may have more white or Amerindian than black ancestors. This American practice, the most extreme form of which is the “one drop rule,” does not alter the objective existence of the African, European and Amerindian racial groups.

Virtually no non-whites take the position that race is an illusion. They have too healthy a sense of racial identity to accept such an odd notion. It is yet another ridiculous idea whites have talked themselves into as part of their overall race hysteria. Whites take great pride in staking out strange but fashionable positions, and then concocting elaborate justifications for them. The more strikingly they contradict common sense the better. To believe and to profess a self-righteous absurdity requires high virtue and mastery of difficult, mysterious teachings — a combination liberals find irresistible. Other examples would be insisting that blacks are as smart as whites or that diversity is a strength, but the purest form of high mumbo-jumbo is race-does-not-exist.

Cracks are, fortunately, beginning to appear in the façade. Population geneticists increasingly report that people can be unerringly classified by “continental population groups,” and thoughtful readers realize that “continental population group” is only a polite way of saying “race.” Researchers like Bruce Lahn of the University of Chicago will continue to find potentially important gene variants that differ in frequency between racial groups. More medicines will appear that have markedly different effects on different races. More people will buy commercially available DNA tests that determine ancestry, and wonder how it is possible to measure something that does not exist. Commentators like Steve Sailer will continue speaking truthfully about race — even if they refuse to accept the implications of what they are saying. Before long, claims about the alleged non-reality of race will be increasingly met with headshaking, smirks, and outright ridicule.

Even liberals have an interest in grasping reality. As Prof. Edwards warns in his article, “[I]t is a dangerous mistake to premise moral equality of human beings on biological similarity because dissimilarity, once revealed, then becomes an argument for moral inequality.” Liberals have built an entire world view on faulty assumptions and willful blindness. Most will go to their graves with their eyes closed; for the rest, there will be an unpleasant awakening.