What are the main results of our study?
1) We discovered five genetic markers that were associated with same-sex sexual behavior. (READ MORE)
Each marker has a very small effect individually — that is, each contributed very little to a person’s sexual behavior. This is not unusual for complex human outcomes. Common genetic variants (typically defined as variants that appear in at least 1% of the population) often contribute only a tiny amount to the variation in the overall outcome.
It is important to remember that these genetic variants alone do not define someone’s sexual behavior. Behavioral traits, like sexual behavior and orientation, are only partially genetic in nature. They are shaped by hundreds or thousands of genetic variants, each with a very small effect, yet they are also shaped in large part by a person’s environment and life experiences. We can therefore say with confidence that there is neither a single genetic determinant of nor single gene for same-sex sexual behavior or sexual orientation. To the extent that sexuality is influenced by genetics, it is more likely that hundreds or thousands of genetic variants are involved. These variants, together with the environment and experiences, shape outcomes like same-sex sexual behavior.
When we analyze all common genetic markers together, they capture between 8 and 25% of the individual differences in same-sex sexual behavior. These results suggest that more markers will be discovered with larger sample sizes.
2) We found some hints as to what these genetic variants do biologically (READ MORE)
It is often hard to say how a genetic marker ties back to the actual biology of a trait; in fact, this can be one of the most challenging tasks in human genetics. But we did find two interesting connections. One of the markers we identified is also associated with balding, suggesting that sex hormone regulation may be involved in the biology of same-sex sexual behavior.
Another is related to our sense of smell. This is interesting, because while odors are important for sexual attraction, we don’t yet understand how this might be related to sexual behavior, and so this finding provides scientists a starting place for follow-up work.
3) Using genetic data, we found evidence that sexual behavior is a highly complex trait and that there is not a single dimension of sexuality (READ MORE)
We found that the genetic influences that contribute to the chance of having sex with someone of the same sex are largely distinct from the genetic influences that contribute to the degree of same-sex sexual behavior (having mostly opposite-sex partners vs. having exclusively same-sex partners). That is to say, the genetics suggest that it is an oversimplification to assume that the more someone is attracted to the same sex, the less they are attracted to the opposite sex. Our findings call into question the validity of single continuum measures like the Kinsey scale.
4) We saw that many of the same genetic markers influence same-sex sexual behavior in females and males, but we also found some markers with sex-specific effects. (READ MORE)
40% of the genetic influences on same-sex sexual behavior were shared by both sexes, and approximately 60% were unique for each sex. This may reflect differences in the biology of same-sex sexual behavior in females and males or may be the result of the different gendered and socialized contexts for expressing sexuality for females and males. However, this is a noteworthy finding, as the genetic overlap between females and males for most traits is usually much higher than this.
5) We found that the genes that play a role in same-sex sexual behavior partly overlap with those for several other traits, including openness to experience and risk-taking behavior (READ MORE)
We also found genetic overlap with some health-related behaviors (e.g., smoking, cannabis use) and with risks for certain psychiatric conditions (schizophrenia, bipolar disorder, and major depression).
While the apparent genetic overlap we observed between same-sex sexual behavior and personality traits is interesting and can help scientists form and explore new hypotheses, we need to be cautious about how we interpret these results, as the social environment likely has much to do with the overlap we observe.
There are reasons why same-sex sexual behavior and other traits may share genetic markers. For example, it could be that one trait leads to another through the influences of environmental factors. For instance, a member of the LGBTQ community may experience prejudice and discrimination based on their sexual orientation and behavior, which would increase their risk for depression. In that case, what might appear to be a genetic association is actually one that is driven by the environment.
What are the limitations of our study?
- Self-reporting. All of the information used in our study that was related to same-sex sexual behavior, identity, or orientation was self-reported. Participants did not have to respond to the question items if they were uncomfortable doing so. The surveys we used are confidential and de-identified to protect privacy and encourage honest responses. Despite this, we expect that some proportion of the responses may be misreported.
- Sex vs. gender. Throughout this FAQ and in our study, we use the terms “female” and “male”, rather than “woman” and “man.” The research here is therefore primarily discussing those who were assigned as female or male at birth. Our analyses and results relate to biologically defined sex, and not to gender or gender identity in any way. Further, as is common in genetic analyses, in the UK-Biobank sample we dropped individuals whose biological sex and self-identified sex/gender do not match. This is an important limitation of our analyses, as they do not include transgender persons, intersex persons, and other important persons and groups within the LGBTQ community.
Categories of sexual behavior. The main question asked to UK Biobank participants (our largest sample) was binary (“Have you ever had sex with someone of the same sex?”). We realize that “never/ever” is an oversimplification and that we combine several sexual identities and expressions together in this measure. To address this issue, we also looked at other indicators of sexual behavior (e.g., the proportion of same-sex partners to total number of partners) and at some additional questions in the 23andMe database related to sexual identity, sexual attraction, and sexual fantasies. Nevertheless, our measures all remain imperfect and cannot in any way make claims about the full complexity or human sexual behavior or sexual orientation (READ MORE)Categories of sexual behavior. The main question asked to UK Biobank participants (our largest sample) was binary (“Have you ever had sex with someone of the same sex?”). We realize that “never/ever” is an oversimplification and that we combine several sexual identities and expressions together in this measure. To address this issue, we also looked at other indicators of sexual behavior (e.g., the proportion of same-sex partners to total number of partners) and at some additional questions in the 23andMe database related to sexual identity, sexual attraction, and sexual fantasies. Nevertheless, our measures all remain imperfect and cannot in any way make claims about the full complexity or human sexual behavior or sexual orientation. For a study of complex trait genetics, where we are looking at millions of genetic markers each with small effects, we need a large sample size; for such large groups, detailed information about individuals’ sexual behaviors often isn’t available. Further, fragmenting data into smaller subgroups could reduce the ability to detect these small effects. This study does reveal a very strong relationship between participants’ genetic data, their answers to questions about sexual behaviors, and the rough proxies for orientation, though the measurement of our outcome remains limited.
The findings are based on statistical patterns in the data as a whole, and no conclusions can be drawn for any particular individual. It is not possible to predict or identify someone’s sexual behavior or sexual orientation from their DNA, nor was doing so our intention. These results cannot be practically used to predict someone’s sexual behavior, orientation, or identity. Because there are so many small genetic effects, and we can’t measure any of them precisely, a genetic predictor is very “noisy”, and knowing someone’s genetic information allows us to guess their sexual behavior just about as well as guessing with no genetic information at all (READ MORE)The findings are based on statistical patterns in the data as a whole, and no conclusions can be drawn for any particular individual. It is not possible to predict or identify someone’s sexual behavior or sexual orientation from their DNA, nor was doing so our intention. These results cannot be practically used to predict someone’s sexual behavior, orientation, or identity. Because there are so many small genetic effects, and we can’t measure any of them precisely, a genetic predictor is very “noisy”, and knowing someone’s genetic information allows us to guess their sexual behavior just about as well as guessing with no genetic information at all. Further, even if a person is born with some of the genetic variants we discovered in our study, our findings show that genetics is only a component of sexual behavior, and that many other factors influence a person’s sexual behavior and orientation. In addition to environmental factors that might affect sexual development during childhood, there may be non-genetic factors that influence sexual behavior and orientation before birth. For example, others have linked same-sex sexual behavior with having more older brothers and hypothesize a role of prenatal exposure to sex hormones. Suffice it to say that the development of sexual behavior and orientation is extremely complex, and unraveling all the factors involved will require more research. Our study is just one beginning part of this effort.
- Study cohort demographics. We only included individuals of European ancestry. This is a common limitation in genetics research that occurs because most available data are from large cohorts of European individuals, like the UK Biobank. Sample sizes of non-European individuals are too small to be well powered to do similar analyses. In addition, most participants in our largest data source, the UK Biobank, were between 40 and 70 years old. We observed a lower percentage of older adults reporting engaging in same-sex sexual behavior. These age and ancestry biases mean that our sample is not representative of the general population and that our findings may not necessarily apply to people from other ancestries, age groups, or demographic categories.
- Nongenetic factors are also important. Though our study focuses on genetic influences on same-sex sexual behavior, previous research indicates that other, non-genetic influences are also important. Our genetic findings in no way preclude the additional influences of culture, society, family, or individual experiences, or of non-genetic biological influences, in the development of sexual behavior and orientation.