How Should We Be Thinking About Genetic Studies?

by Kayt Sukel

June 16, 2014


In February 2013, the New York Times Magazine published, "Why Can Some Kids Handle Pressure While Others Fall Apart?" The article, which quickly went viral, focused on school stress, and placed great emphasis on Catechol-O-methyl transference (COMT), a gene that codes for an enzyme responsible for breaking down neurotransmitters in the brain. The difference between a kid who can't handle the pressure, a "worrier," and a kid who takes it all in stride, a "warrior," the authors argued, may come down to how the COMT is expressed. It was a compelling story-and one that parents seemed to cotton to (likely in hopes of explaining their own kids' school performance). But many experts argue that the science is not that simple: there are important caveats that should be cited when discussing genetic studies, particularly genome-wide association studies.

In the past decade, the field of behavioral genetics, or the study of how genetic variants influence human behavior, has grown dramatically. Researchers no longer just look at how our DNA affects our health but also our cognition, intelligence, and personality. As the number of these studies grow, it is easy to want to accept simple explanations like the "worrier/warrior" one. Not to mention studies that discuss "monogamy," "divorce," "Mommy," and "empathy" genes. Such folk science stories seem, based on our own experience, to fit. We tend to like simple, causal explanations when it comes to science. Yet, such oversimplifications do not give us the full story on how genes really influence human behavior. So how can we think critically about genetic studies that concern behavior and the brain? What factors should we consider before determining the true impact of a gene on human behavior?

The Single-Gene and Small N Problems
Single-gene stories are a staple of early science education. Some time after discussing Gregor Mendel, pea plants, and inherited traits, students often learn about diseases like Huntington's disease and Fragile X syndrome, both of which are caused by mutations in a single gene. Certainly, early genetic studies were based on an idea that even complex traits were linked to a single gene defect. Subsequent work, however, has shown that notion to be false.

"One of the things that we've learned is that individual genetic variants seem to have very small effects on complex traits," says Christopher Chabris, Ph.D., a psychologist at Union College. "The general lesson is that human behavioral traits are influenced by hundreds or perhaps even thousands of common genetic variants. So a single gene variant explains very little of the variability we see in traits like empathy or risk-taking. And that simply wasn't known 20, or even 10 years ago."

So much for single-gene explanations! And given that most traits are the result of polygenic, or multi-gene, contributions, Chabris argues, studies need sample sizes in the tens or probably hundreds of thousands to separate the genetic wheat from the chaff.

"If multiple genes are involved, and each individual gene's effect is tiny, then you have no hope of finding anything other than noise if you have 70 participants or 170 participants or even 700 participants," he says. "7,000 might not even be enough-although that's starting to approach the number you need. But the point is that you need a lot of participants, thousands upon thousands, to find these kinds of effects. And it's a real challenge."

The numbers matter. The COMT study cited by the "warrior/worrier" article looked at 779 students. The majority of genetic studies are published with n's in the hundreds or thousands-but some have been published with a sample sizes as low as 23. So Chabris cautions that genetic studies with lower n's should be taken with a grain of salt. "Those studies may be false-positives. With larger samples, the researchers may find that the results they saw simply aren't there anymore," he says.

The Correlation/Causality and Gene/Environment Interaction Problems
It would be all too easy to think about single-gene explanations as deterministic-certainly, that is the way they are often described in the press. If you have one variant of the COMT gene, you will be a "warrior," all but impervious to stress. If you have the other, you are a "worrier," and will be felled by pressure, no matter the task. But genetic studies are correlational by nature. They cannot discover any causal relationships.

"Suppose you have the particular gene variant. So maybe you're one or two percent more likely than someone who doesn't have the variant to be able to deal with stress or whatever trait that paper is talking about," says Chabris. "But it's not deterministic. Even if you have a particular variant, there's no guarantee that you're going to have the phenotype."

It's important to remember that genetic studies are probabilistic, not deterministic. A particular gene-linked behavior that comes out of a group study may not apply to an individual. Yet there are factors that can up the probability that an individual with a particular genetic variant may exhibit a certain phenotype. And those factors are not biological in nature-they are factors experienced in the environment. Poverty, abuse, and education level, for example, also play an important role in the development of particular behaviors. MAOA is often discussed as another type of "warrior" gene, due to its links to violent behavior. But, of course, its actual function has little to do with criminal enterprise. MAOA is a gene that codes for an enzyme responsible for degrading certain types of neurotransmitters in the brain.

In a talk about neuroscience and the law at Neuroscience 2013, Nita Farahany, director of the Duke Institute for Genome Sciences and Policy, discussed how the low expression of MAOA crossed with childhood maltreatment was linked to violent and anti-social behavior later in life. "It shouldn't surprise you too much in looking at this data that this research has come into a number of criminal cases," she informed the audience. But her key point was that violence was as much linked to the environment as it was to the gene-and it can be difficult to tease the individual contributions of the two apart.

The Need for Genetic Literacy
Patricia Churchland, B. Phil., a philosopher and neuroethicist from the University of California San Diego, as well as a member of the Dana Alliance for Brain Initiatives, says that interpreting the results of genetic studies requires thinking like a skeptic and a critic. "You have to put the hype in perspective. Because as much as people want to say that there's a gene for this and a gene for that, often the data are just an association within a limited population," she says. "And that, unfortunately, doesn't always tell us that much about what a gene does."

Both Churchland and Farahany call for more responsible communication by scientists and journalists when it comes to genetic research.

"We need to explain the context of genetics in a very different way now that the technology and science has improved," says Farahany. "Now more than ever we must help people understand there's no such thing as a single gene for any behavior, but a complex set of genetic, epigenetic, and environmental factors that influence behaviors."

Huda Zoghbi, a molecular geneticist who studies rare diseases at the Baylor College of Medicine, says that education is also key. She suggests that science educators could go a little deeper when discussing genetic concepts with students, even as early as elementary school.

"Certainly we should still highlight single gene defects. But it is also important to explain that there are many genes involved in behavior," she says. "Today, we don't cover the intricate relationships between genetics and experiences, perhaps, as much as we need to. We should emphasize that the genome is the framework for how genes affect health and behavior. But we should also discuss that there are many other factors that modulate those effects even in a healthy genome."

Churchland concurs. "It won't be easy, I know. We are all predisposed to want simple explanations-simple, causal explanations," she says. "We need to help people understand that genes essentially never act alone, and a single gene linked to a single trait is highly unusual. Any behavior involves a very complex network of genes and dynamical interactions between those genes and the environment. It's an awesomely tricky business, certainly. But it's a reminder of just how terribly important scientific literacy is.  Every year, with the launch of many new kinds of studies and techniques for getting meaningful data, scientific literacy matters ever more."

Published June 2014

 Yeh T-K, Chang C-Y, Hu C-Y, Yeh T-C and Lin M-Y. Association of catechol-O-methyltransferase (COMT) polymorphism and academic achievement in a Chinese cohort. Brain and Cognition. 2009, 71(3): 300-305.

 Shetty P. Monogamy gene found in people. New Scientist. 1 September 2008.

 MacFarlane J. Divorce? It could be in your genes: How DNA could play a big part in how much people argue. The Daily Mail. 25 February 2012.

 Rosoff PM. In search of the Mommy gene: Truth and consequences in behavioral genetics. Science, Technology and Human Values. 2010, 35(2): 200-243.

 MacMillan H. Is empathy in our genes? Health. 15 November 2011.

 Hamer, D. Rethinking behavior genetics. Science. 2002, 298(5591): 71-72.

 Kogan A, Saslow LR, Impett EA, Oveis C, Keltner D and Saturn SR. Thin-slicing study of the oxytocin receptor (OXTR) gene and the evaluation and expression of the prosocial disposition. Proceedings of the National Academy of Sciences. 2011, 108(48): 19189-19192.

 Jacbonson, K. Considering Interactions between genes, environments, biology and social context. Psychological Science Agenda. April 2009.

Caspi A, McClay J, Moffitt TE, Mill J, Martin J, Craig IW, Taylor A and Poulton R. Role of genotype in the cycle of violence in maltreated children. Science. 2002, 297(5582): 851-854.