By Kenneth Weiss

Ever since Darwin, there has been great controversy over the degree to which genes determine who we are. The reason is that if genes determine individuals' or groups' nature, value judgments can be assigned to those genes-and the people who carry them-and social policy can be implemented accordingly. Much good has come from the pursuit of genes that cause disease, but racism and eugenics based on uncritical genetic determinism have also visited incalculable harm upon people. 

There's no legitimate doubt about the importance of genes in the making of organisms, or of the role of genetic variation among them. The theory of evolution led us to understand that life has diverged from a common origin, by way of inherited variation that has been screened by luck and its bearers' ability to reproduce. 150 years of research have revealed much about the nature and importance of genes, and today the view of life that Darwin spawned holds genes to be primary instruments of the development of organisms.

Organisms develop from single cells. There is more in a cell than DNA, but only DNA carries the specific information that controls biological function. In a mechanistic sense, DNA is ultimately responsible for the structures of individuals. Differences in traits among individuals are affected by differences in DNA, and by extension, populations that are isolated from each other gradually accumulate genome-wide differences. These differences affect both functional and non-functional sites in the genome, some of which are harmful.

It may thus seem unexceptionable to look for harmful genes and hope to intervene. The idea begins with detection, treatment and prevention of readily identifiable genetic diseases. That is now routine for some diseases. Hospitals screen newborn infants for a number of genetic conditions, and many couples submit to DNA tests to determine the possibility that they might pass a condition on to their children. The hope is that harmful alleles may even be removed from the human gene pool. Today such measures are largely voluntary, though screening newborns for some genetic disorders, notably PKU, has been mandatory for many years in the U.S., and widely accepted because there is effective treatment.  

Things become problematic when "public good" is used to advocate mandatory social policy. The belief that all traits are meaningfully genetic opens the extension of policy concerns beyond clear-­cut genetic disease. Historically, that has stressed behaviors that experts and policymakers have chosen to define as undesirable, with special attention given to traits like homosexuality, insanity, drug or alcohol addiction, immorality, or aggression. Historically, that's when the wheels came off.

Post World War II revulsion over abuses made in the name of eugenics and its racist nadir in Nazism led the nature-nurture cycle to swing to an environment-centered view of human traits, accompanied by declarations of human rights to prevent such kinds of discrimination. Physicians were even advised to avoid specializing in genetics because genetic diseases were rare and nothing could be done about them. It was felt that the common diseases and the diseases of older age were caused by environmental exposures and lifestyles; they were not genetic and could be addressed by therapeutic or preventive measures. Similarly, psychologists and social scientists took behavioral traits off the genetic table.

But there has been a resurgence of genetic determinism. Modern genetic technologies have made it increasingly possible to document the structure of DNA and to characterize gene function. Along with this came increased ability to search for genetic effects underlying disease. In the process, awareness of raw abuses of the past has faded into a rather stunning complacency in the belief (once again) that our actions are entirely value-­neutral and benevolent.

The fact is that most clearly measurable traits show substantial heritability. Offspring resemble their parents, even when environmental and statistical errors are accounted for. It is typical for roughly 30-50% of variation in a trait to be statistically accounted for in terms of genetic variation. Despite well known problems with heritability as a measure, this clear cut familial correlation feeds the determination to identify the genetic basis of any disease. There is considerable controversy about how efficacious such gene 'mapping' approaches are, but nobody wants to stand in the way of prevention and treatment of disease.

If disease traits that everyone agrees should be addressed clearly show genetic influences, then there is no scientific reason to think that other traits, including even culturally sensitive behavioral traits or group differences, should not also have some genetic basis as well. To deny that genetics is a part of the causal spectrum of inconvenient traits is to be an ostrich to the facts.

This may be difficult to accept in light of the brutal history of abuses in the name of Darwin, Mendel, the inherent value assigned to persons on account of their genes, and the patently facile way in which tiny indications of difference are extrapolated across vast evolutionary time periods. Reconciling the legitimate scientific facts with the legitimate social fears is not easy.

Meanwhile, researchers are now excitedly proclaiming they have found the genes 'for' all sorts of behavioral traits, including such things as athletic or musical ability. Unfortunately, much of the stress is on the same kinds of "undesirable" traits that led to the abuses of the eugenic era. This is a return to the acceptance of inherency ushered in with the Darwinian view of life. As in the earlier reasoning, no one wants people to suffer needlessly from genes predisposing them to disease, and no one wants society to suffer needlessly from genes predisposing an individual to harmful social behavior. And as before, a belief in genetic inherency labels people from birth. It is entirely predictable that genetic screening that identifies people with a genotype supposedly predisposing them to crime will lead to various forms of surveillance (or mandatory preventive pharmaceutical "therapy") before they commit their predestined offense.   It is clear that this is only statistical risk, usually weak, and most of the individuals targeted as having the "criminal gene" will never commit an offense.

It is not easy to decide how to strike a proper balance between genetic determinism and civil society. At least, we should take a closer look at the facts. The issues are subtle and perhaps that is why they are not easily absorbed, even by professionals. But they are important.


People and groups will always differ, on anything one can measure

No two people or populations will be exactly identical on any measure. There will always be some difference. Whether it is statistically significant depends on the sample, the precision of measurement, and subjectively chosen probability criteria (like the typical 5% level). No one can pretend that such differences cannot, in principle, affect any trait. This is why if one chooses to sample from different continents one can make "races" seem so real, even though we know that the evolutionary processes are more continuous in nature.

However, one has to be careful about what such differences are. Samples are usually compared by mean values, like average IQ scores or blood pressure levels. But usually there is a large overlap in the distributions, as illustrated in the Figure where the distributions differ by a typical half-standard deviation. The difference in the mean may be statistically "significant;" but that is not the same as importance, or even the main story. Similar statistical issues apply to comparisons of individuals defined as affected versus unaffected by disease or behavioral traits. Whether the mean differences are important relative to the overlap is a societal judgment. Such judgments have been, and are still, widely used to support racist views.


Why is genetic causation so elusive?

Mapping studies have been done on almost any trait one can imagine and in every major research organism, from bacteria to human. They have found very similar results that are entirely consistent with the genetic effects expected from our understanding of the evolutionary process and the nature of the way genes work.

Generally, there are one or a few strong effects at individually identifiable locations in the genome. The effects of these sites achieve at least suggestive statistical significance according to standard (though subjective) criteria, and have replicable association with the presence of the trait. However, such sites usually account for only a small fraction of the estimated heritability of the trait. Accompanying them, ever-larger studies have shown that a host of sites-often estimated in the hundreds or more-have individually minute effects. Even in aggregate, these minor sites, or "polygenes," still only account for a small fraction of the heritability. They rarely are replicable among studies or populations.


How predictive are genes?

Such findings reinforce the idea that genetic variation contributes overall to the trait in question, but does not directly address the question of the predictive power of individual genotypes, which is crucial to an understanding of genetic causation. The same mapping results unambiguously show that the non-­genetic environmental effects account for the bulk of variation.

Furthermore, heritability implies genetic involvement but says little or nothing about genotype-specific prediction. Individual multilocus genotypes are unique, and therefore we usually have little way to predict their specific effects because statistical association depends on replication. More fundamentally (or even fatally) serious is that environmental exposures contribute the bulk of effects but are both poorly understood and unknowably changeable, meaning that the causal effects of all but the few, usually rare, strong genetic variants are contextually relative and quantitatively unpredictable.

The same is even truer of behavioral traits. For behavioral traits, much depends even on the purely cultural nature of their definition. As an obvious example, it is those in charge who define what is "antisocial" or criminal behavior. Unscrupulous hedge fund managers walk the streets freely, while a large fraction of the US criminal population is incarcerated for drug-related offenses. A search for criminality genes would compare these "normal" and "criminal" groups to find genetic differences between them-genes "for" criminality. But if drugs were legalized, what had been offenders' genotypes would overnight become "normal." This cultural component is so obvious it is a wonder anybody accedes to the kinds of studies searching for genes "for" behavior increasingly being funded in our gene-centered age.

One must acknowledge that there is, at present, none of the racist or vitriolic judgmental fervor that drove the classic eugenic movement. But the genetic ideology and target traits are similar, and that's what raises caution about where the enthusiastic promises of finding the genes conferring specific value-laden behaviors may go.

What, if anything, can be done to recognize the effects of genes and yet protect against exuberant application of genetic determinism? Could one identify a line beyond which research is banned because its findings might be too likely to be misused? That may seem like unacceptable or unenforceable censorship, but in fact, society routinely makes such decisions. For example, investigators are not allowed to torture people as part of a scientific experiment (even if, for some reason, they could be persuaded to consent).

We already make informal discriminatory decisions based on traits that have some genetic basis. Intelligence, however you want to measure it, surely is in that category. Bright students are assigned faster academic tracks, are admitted to the best universities, or are given special mentoring and encouragement. It seems acceptable to base such discrimination on test scores, perhaps because that at least reflects achieved ability. But many would object if we used genetic tests instead, even if, for example, it might identify those in whom even earlier investment was warranted.

The search for simple answers affects society and science alike. The complexity and uncertainties of nature encourage both progress and problems, but the lesson is clear: Genes are involved in everything, but not everything is "genetic."

Kenneth M. Weiss, PhD, is Evan Pugh Professor of Anthropology and Genetics at Penn State University.

(The author thanks Anne Buchanan for helpful critique of this commentary.)

Search: GeneWatch
The Gene Myths series features incisive, succinct articles by leading scientists disputing the exaggerations and misrepresentations of the power of genes.
View Project
Biowarfare and BioLab Safety
View Project