A recent study has drawn controversy by implying genetic links between bisexuality in men and a propensity for risk-taking. This research on human sexual behavior, published in January in Science Advances, is an example of a genome-wide association study (GWAS). Such studies compare entire genome sequences from many people in a search for areas of overlap between genes and certain traits. The authors of the new study report that bisexual behavior in men is genetically distinct from exclusively same-sex behavior and suggest that the genes underpinning bisexual behavior are also linked to possessing an inclination for risk-taking and to having more children.
“The basic finding is that bisexual behavior and number of children are genetically positively correlated,” says senior study author Jianzhi Zhang, a professor of ecology and evolutionary biology at the University of Michigan. The study found that certain gene variants, or alleles, were more common in men with self-reported bisexual behavior. These same alleles were also associated with a self-reported inclination for risk-taking. Some men who reported exclusively opposite-sex behavior also carried the gene variants associated with bisexuality and risk-taking behavior; in these men, these genes were associated with reporting a higher number of sexual partners. In premodern societies, the study notes there was a strong correlation between higher number of sexual partners and fathering more children. The researchers posited that these correlations could be an evolutionary explanation for why bisexual alleles and behavior persist in the human population.
Zhang and his co-author based the findings on data on about 450,000 people from the UK Biobank, a database of participants’ genetic information and self-reported answers to survey questions that is commonly used in GWASs. Starting in 2006, UK Biobank recruited participants who lived in the U.K. and were between 40 and 69 years old. The new study, like many GWASs, only included participants who identified as “white,” and relied on categories of bisexual behavior, risk-taking behavior and number of children that reflected just a few of those survey questions. The “risk-taking” trait was based on a single question from those surveys: “Would you describe yourself as someone who takes risks?”
On supporting science journalism
If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
The paper has prompted critique and concern from some geneticists and other human sexuality researchers. They have pointed out that the method behind the new work and other GWASs has important limitations. These researchers have also cautioned that misinterpretation of GWASs by scientists, the media and the public can cause harm.
The notion of a single “gay gene” that accounts for homosexuality goes back to a widely reported on 1993 study of 114 families that linked genetic markers on the X chromosome to male sexual orientation. Subsequent studies and scientific critiques have since cast considerable doubt on this idea, and many genetics experts consider it debunked.
This doesn’t mean genetic studies of human behavior are inherently incorrect. Yet when people, particularly marginalized people, are the subjects of that process, special care should be taken in how research is conducted and interpreted, says Steven Reilly, as assistant professor of genetics at Yale University.
In Reilly’s view, the new bisexuality study could have been conducted and interpreted with greater care, and others agree. “It is a paper that has technically correct correlational analyses in it,” says Robbee Wedow, an assistant professor of sociology and data science at Purdue University. Beyond the basics, however, Wedow says, many of the new study’s stronger claims are weakly supported, with “overhyped results.” Joanna Wuest, an assistant professor of politics at Mount Holyoke College, who studies gender and sexuality politics, echoes this sentiment. “It’s a lot of storytelling based around not a lot of data,” she says.
Zhang acknowledges that his study had limitations and relied on some assumptions but denies that its results were weak or that its conclusions were too strong. Many of the critiques of the paper aren’t just about Zhang’s work, however. They also get at larger limitations of contemporary genetic research and the way it’s interpreted by scientists and the public.
Here is what scientists generally agree on: Twin studies and other work indicate that some amount of sexual orientation is heritable. The level of heritability found by twin studies varies, but it is under 50 percent in nearly all cases—meaning research suggests at least half of the factors that determine sexual identity are social, cultural and environmental—not passed down through DNA. Advances in DNA sequencing techniques and GWASs have enabled a more recent wave of research into the genetic basis of human same-sex behavior. The most prominent of these is a 2019 GWAS paper in Science co-authored by Wedow that found that the genetic underpinnings of same-sex behavior are complicated—with a wide variety of genes contributing to somewhere between 8 and 25 percent of sexual orientation.
Yet the finer points of the 2019 study and subsequent papers are often missed. For one, GWASs can only suggest associations, as implied by the name. They can’t tell you the underlying factors that lead to those associations, Reilly and others say. “I do believe that GWASs show you much less than some people think they do,” says Agustín Fuentes, a professor of anthropology at Princeton University, who studies human evolution and behavior. “GWASs are a great statistical tool to find patterns and trends across large and complicated datasets, but they tell you nothing about cause and effect.”
In their recent study, Zhang and his colleague did not explicitly state genetic causes of bisexuality, risk-taking behavior or increased number of children. “We never [said] in the paper that risk-taking causes bisexual behavior or bisexuality causes risk-taking. What we [said was] that the genetic underpinnings of risk-taking and the genetic underpinnings of bisexuality are overlapping,” Zhang explains.
The correlations the researchers drew from the GWAS findings implied possible causation, Fuentes says. He views this as an overreach because, he says, the initial correlations were tenuous to begin with.
One limitation of these studies stems from the way that genetic researchers define behavioral categories. Sexuality is a difficult trait to measure—and most metrics are poor, which, Wedow says, was a limitation of both the new study and his team’s 2019 work. Because sexual orientation can be hard to pin down, scientists rely on proxies.
Zhang’s study defined bisexual behavior as self-reported sex with people of both the same and opposite sex. This type of categorization doesn’t capture the idea that someone has the autonomy to self-define, Wuest says. It can also exclude many people’s lived experiences, Fuentes says. “The actual category of bisexuality is an incredibly complicated one. The psychological and behavioral stuff associated with it does not necessarily mean [a person] has had sex with someone with a penis and someone with a vagina,” he notes.
Stigma and circumstance can prevent many people from engaging in same-sex behavior or reporting it to researchers, Reilly says. This could have been especially true of the older age cohort that was included in the UK Biobank data. Same-sex acts between men were a criminal offense in Scotland and Northern Ireland until the early 1980s. And people who identify as gay or experience exclusively same-sex attraction may face sociocultural pressure to engage in opposite-sex relationships even if they aren’t bisexual.
Self-reporting is also deeply flawed, Fuentes says. He notes that qualitative studies suggest respondents often lie when responding to such surveys. Even if people believe they’re being truthful, definitions of sexual activity vary, he adds. One person’s clear, affirmative answer could be another person’s “Well, that wasn’t exactly sex.” And one person might count experiences from childhood or youth in a self-concept of sexuality while another might exclude them, Fuentes notes. The definition of risk-taking is also deeply subjective. And when it comes to the number of children a person has, some proportion of men may not know the actual count, Fuentes adds.
There are also inherent limits to the genetic links highlighted by association studies. In GWASs, each bit of DNA that is found to correlate with a trait usually only accounts for a very small portion of that trait’s presence—on the order of a couple of percent or less, Reilly says.
Together, all these small gene effects add up to a total effect that is still small. GWASs don’t account for the nongenetic factors that nearly all studies suggest play a bigger role in human sexuality than genes alone. Analysis from the 2019 GWAS shows that birth year is a factor in whether or not people report same-sex behavior. U.K. Biobank participants born in 1970 reported about three times as much same-sex behavior than those born in 1940, Reilly and Wedow both note. “I would not think that there is a biological or genetic reason” for that, Reilly says. “It shows that there’s a lot of strong other societal things at play here.”
Critics of human genetic studies often point to how they can easily be misused. Lots of GWASs aim to assess complex traits such as academic achievement and IQ. In some cases, bad actors have tried to willfully twist genetic research to support racist ideology. It’s not too far of a reach to imagine that studies of sexual orientation, too, could be fuel for harmful pseudoscience, Reilly says. After the 2019 study by Wedow and his colleagues was published, despite the authors’ many good faith communication efforts, a misguided app was released that claimed it could tell individuals “how gay” they were based on their genetic data. The app was taken down after efforts by authors of the study and others, but Reilly believes the false idea of a genetic test for sexuality could linger.
Other damage might be less obvious, especially when cultural assumptions come into play. The idea that bisexual people are risk-takers could be interpreted as suggesting they are promiscuous—a common and harmful stereotype, Wuest says. In at least one instance in the past, she adds, this belief led to serious social consequences: during the early years of the HIV/AIDS epidemic, bisexual people were scapegoated as viral vectors.
“In a perfect world, there are no harms” from studying the genetics behind traits like sexuality, Reilly says. “Knowing this information is great. It’s interesting. It excites some people’s understanding or helps them understand themselves better.” But he emphasizes that reality isn’t perfect.