go home Genetics

April 2004

Genetic Variation

Mike Rienzi

The question of intra-group vs. inter-group genetic variation requires some more discussion, not only for the “is race real?” debate, but also because Frank Salter is careful to specifically discuss inter-group differences in the frequencies of distinctive genes in his work on “ethnic genetic interests.”

It is true that for the vast majority of gene loci, there are no population-specific differences in gene frequencies while, at the same time, there is considerable random variation in types of alleles (gene frequencies) at these loci in the overall human population. From these two facts derives the obvious: there must be more genetic variation within human groups than between groups. After all, let us carefully consider what is being said. First, we state from the outset that for most of the human genome there are no population-specific differences in gene frequencies. But then we say there is considerable variation among humans as a whole. Thus, it follows that if this variation exists in toto, but is not found between groups, then it must be found within groups, as it indeed is.

It also follows that if the genetic variation for most of the genome is randomly distributed, it is of course possible to find an Irishman who is, overall, genetically more similar, by mere chance, to a Nigerian than to another Irishman. This is because the first Irishman and the Nigerian may, only by random chance, share more similarity in these continuously distributed non-population specific alleles than does either to the second Irishman.

However, there is a minority of gene loci for which population-specific differences in gene frequencies exist. For many of these, the differences in gene frequencies are small, but for some, the differences can be quite large. Obviously, genes that encode for phenotypic differences between population groups are included in this group, but there are many markers that do so as well. Such markers are used in the AncestrybyDNA test, which can determine majority racial identity with ~100% accuracy and minor admixture with up to 95% accuracy dependent upon if the estimate of admixture is above the levels known to be statistically significant. The existence of these population-specific loci is what accounts for that % of human variation that exists between groups.

However, since most of the genome does not exhibit population-specificity in variation, intra-group variation is significantly higher than inter-group variation. As Sarich and Miele point out in their book on race, since humans are diploid, a single individual can contain considerable genetic variation himself/herself; thus a large fraction of the intra-group variation is simply found within individuals (and can be found within families as well). But still, the point remains that intra inter group genetic variation, and it is possible for two members of different races to share more similarity in their overall (note: overall!) genome than with co-racialists. Before discussing the implications (or lack thereof) of this, I’d like to give an extremely simplified analogy to assist in making this clearer.

Let’s use marbles instead of gene loci and different colors for these marbles in place of variable alleles. For different population groups, we’ll use groups of boys and girls, each member of each group possessing 20 marbles. The population of the entire school in which these children reside is analogous to the overall human population.

Let us assume the marbles for each child are numbered 1-20 and can come in one of several colors; e.g., blue, pink, red, orange, yellow, green, purple, etc. Let us provide the following rule. For marbles 1-18 for each child, the marble can be of any of the colors, and this is randomly distributed. There are no differences between the boy population and the girl population in the frequency of colors in the set of marbles 1-18. There is considerable variation in color for the school as a whole: looking at ALL the children lumped together there is an enormous variability in the combinations of colors any one child can have for #1-18. But again: no differences between boys and girls in the color frequencies in the first 18 marbles. It is very possible for a given boy to have a set of marbles # 1-18 more similar to the color combination of a girl than to that of another boy.

But there is another very important rule. For the 19th marble, there is a 78% possibility that the marble is blue if owned by a boy, while there is a 84% possibility that the 19th marble is pink if owned by a girl, and there is a 94% chance that the 20th marble is red if owned by a boy, and a 90% chance that it is yellow if owned by a girl.

Now what do we have? There is still more color variation within groups than between groups (90% of marbles exhibit variation which is random between groups). Looking at the overall color distribution of marbles, it is still possible to find boys more similar to girls than to other boys. However, there are now distinct differences between the two groups, as exhibited by the color distribution of two out of the twenty marbles. Can we not begin to separate the two groups based on this? Those marbles that differ in color frequency in a population-specific way can be called “distinctive marbles”, just as genes that differ in allele frequency in a population-specific manner can be termed “distinctive genes.”

Another point. If all the boys were removed from the school, what would happen to the color frequencies? For marbles #1-18, the frequencies would remain essentially unchanged, because, as we stated, there are no population-specific differences for those marbles, marble color is distributed randomly between boys and girls for that set. However, there would be an extremely marked change in the color frequency for marbles #19 and 20 if the boys were removed: blue-colored #19 marbles and red-colored #20 marbles would sharply decrease in relative frequency, while pink and yellow colored 19s and 20s would increase in relative frequency. And what if the color of marbles was an important competitive trait?

Going back to human genes, the high level of random intra-group genetic variation is not important for group genetic interests, precisely because this variation is both ubiquitous and random. Gene frequencies that do not differ between groups - gene loci that show randomness of alleles between groups -- these are not the stuff upon which inter-group competition and genetic interests are built. For most of the human genome, it would not matter if all humans except for Bushmen were killed, because the random variation in gene frequencies within Bushmen is at least as great as that of any other human population. Nothing is lost if the distribution of these gene frequencies is similar for all populations. However, for those distinctive genes, for which there is considerable inter-group variability in gene frequencies, much would indeed be lost if specific populations are diminished or eliminated. Members of groups thus have interests (“ethnic genetic interests”) in the frequencies of those distinctive genes, which is what in fact distinguishes one human group from another at the genetic level. The randomly distributed genes are found at ~equal frequencies in every group and are therefore not a matter of group “concern.”

Let us assume the commonly accepted values of 85% genetic variation within populations and only 15% variation between populations (I realize that Sarich and Miele make a reasonable argument for a much smaller intra-group value than 85%, but for the sake of this argument, to show the power of the argument, let us use the race deniers’ high figure of 85%). If all people on Earth except Bushmen became extinct, 85% of the genetic variation found in the human race would still exist within those Bushmen. However, the loss of any specific human population would result in the loss of that 15% of human genetic variation that is inherent in the genetic differences of that group with all other groups.

Rare, unique, and distinctive objects, which if lost are gone forever, are by their nature far more valuable than ubiquitous objects distributed everywhere. A rare flawless diamond is more valuable than a common rock. Furthermore, those genes and gene frequencies that result in the phenotypic differences between population groups are concentrated in that 15% of variation between groups. Therefore, not only is this lesser level of variation more valuable because of its uniqueness but also because of its utility. Thus, by analogy, a rock containing uranium is not only far more rare and more unique than any common rock, but the potential usefulness of this special uranium-containing rock is infinitely greater.

Thus, at the genetic level it is these population-specific differences in gene frequencies that define race. That is why it is not important that an Irishman may share more similarity in his overall genome to a Nigerian than to another Irishman, for that similarity is based only upon those loci which exhibit random variation of alleles within and between all populations. However, the Irishman in question will in fact share more similarity with another Irishman than with a Nigerian for those gene frequencies that are not randomly distributed, but exhibit population-specific patterns.

This is why it would be adaptive for an Irishman to support his fellow Irishman than a Nigerian who may be more similar in the entire genome. If all Nigerians became extinct, this would not change the worldwide frequencies of those randomly distributed genes. They will still be found in similar frequencies in all other populations. However, if all Irishman except for our Irish person of interest became extinct, then there certainly would be a diminishment in the “Irish-specific” gene frequency profile.

Thus, when comparing two people as regards genetic similarity, what are important are those genes that vary in a meaningful (i.e., non-random) way between population groups. Another analogy would be if someone was trying to send you an important message over the radio, but this faint message was being obscured by loud, random static. The static is louder than the message, but it is the message -- a unique piece of useful information - that is important.

Granted, intra-population genetic variation is important in its specific context, but, in relative comparison to inter-population differences, from the specific perspectives of “genetic interests” or phenotypic variability, it indeed IS noise. We should not let race-deniers deter us from understanding the message of distinctive genes just because they would have us concentrate on the background genetic static.

Furthermore, as people such as Henry Harpending and Frank Salter make clear, the level of genetic variation between populations (whether it be 15% or 12.5%, etc.) is approximately equal to the coefficient of kinship (which is half the coefficient of relationship) between grandparent and grandchild, or between half-siblings. Thus, even from a strictly genetic standpoint, it is the distinctive genes that matter the most.

As regards that PBS series, I’d like to point out my article in American Renaissance on that topic, which Sam Francis has been kind enough to mention in some of his own writings. However, I believe the argument made in the current essay is an important addition to what I, and others, have written about biology and race-denial in the past. In summary, it is the very ubiquity of the intra-population genetic variation that limits its importance, and it is the uniqueness of the inter-population variation (never mind the important phenotypic consequences of this variation) that defines its relative importance.