Homosexuality is widespread among animals, and it’s present in every culture we’ve looked at. All of which suggests that it’s in some way innate to the human species. But figuring out what “innate” means is a real challenge when it comes to a complicated suite of behaviors like sexuality—behaviors that are also heavily influenced by culture.

Studies of twins have suggested that genetics can influence homosexual behavior, accounting for roughly 20% to 30% of its frequency. But attempts to find specific genes that mediate this influence have come up empty. Now, an absolutely enormous study has found a number of genetic sites linked to homosexual behavior. But collectively, they account for a tiny amount of the total genetic effect, and their influence is complicated: different in men and women, and different even across the spectrum of sexual attraction.

Very big data

There are a number of human traits that have a clear genetic influence but aren’t clearly tied to any single gene. Height is a good example. Height is influenced by a large number of genes, each of which on its own has only a tiny effect—there’s no “tall gene.” And, obviously, environmental influences like nutrition are also strong influences on height and can completely swamp the impact of genetics.

While it’s possible to identify genes with very strong influences by tracking their inheritance in a handful of families, this won’t work for identifying any of the weak influences that contribute to human height. Instead, researchers have benefitted from the development of DNA chips, which can survey tens of thousands of individual locations in the genome at once. These have enabled what are called “genome wide association studies,” or GWAS. By taking a large group of individuals and looking at sites across the entire genome, it’s possible to track what parts of the genome are associated with weaker influences.

The weaker the influence, the larger the starting group you need to pull out its effect. For the new study, researchers turned to one of the biggest sources of genetic data available: the UK Biobank, which has genetic data from about half a million UK residents, along with health, demographic, and personal data. Critically for this work, that personal information included self-described sexual preferences. To supplement this, the team turned to the commercial genetics service 23andMe. Tens of thousands of 23andMe customers have agreed to allow their data to be used for research; a large number of them have answered a survey about their sexual practice.

(There seems to be a degree of self-selection at 23andMe, as a higher percentage of those who provided data on their sexual practices identified as homosexual than the percentage in the population as a whole. That shouldn’t affect this specific research question, though.)

All together, the two data sources provided genetic and sexual practice information on over 475,000 individuals. Three smaller groups from separate sources were used to replicate the findings. And the research plan had been pre-registered in order to ensure there wasn’t any post-hoc data analysis done to search for spurious but statistically significant results.

Much ado about not much

One of the nice things about DNA chips is they provide a good estimate of how closely related any two individuals are, independent of issues like adoption and uncertain parentage. By comparing this relatedness to who had same-sex partners, the researchers were able to estimate the total genetic contribution to this behavior: 32%. That suggests the rest is some combination of environmental and social influences. The UK sample, for example, tended to be older, and thus grew up at a time where homosexuality wasn’t accepted—a social influence on behavior that seems to be fading, which could cause the genetic influence to be more prominent in future studies.

In analyzing the entire population, the researchers came up with a grand total of two locations in the genome that were associated with homosexual activity. They also analyzed males and females separately, finding two additional sites that influenced male behavior but not the behavior of females, and one that influenced female behavior but didn’t do anything in males.

Overall the effect of any of these sites was incredibly small. While the analysis of some of the genes is still ongoing, the authors note that men with one of the two male-specific sites were only 0.4% more likely to engage in homosexual activity. And collectively, the authors suggest that the genetic influences they could track via GWAS correlated with a maximum of 25% of homosexual behavior—that’s lower than the 32% figure calculated by relatedness. So, it’s possible that there are also some more complicated genetics yet to be sorted out.

The two genes that the researchers described in this paper (the male-specific loci) are also a bit odd. One seems to be involved in the response to sex hormones, which makes a tremendous amount of sense. But the other seems to be involved in olfaction. While it’s possible to hand-wave a connection there—must be pheromones!—we currently have no idea what the significance of this might be.

The researchers also tested whether the genetic correlations of homosexual behavior correlated with anything else. They came up with a variety of things, like openness to new experiences and cannabis use. But openness to experiences also correlates with cannabis use, so it’s not at all clear how these things might be related; we’re nowhere close to teasing out cause and effect here.

Off the spectrum

The researchers were also able to do an additional test thanks to some of the personal information that was provided. The data allowed them to separate out groups of people who had only had same-sex partners occasionally from those who were nearly exclusively homosexual. Their analysis suggested that, genetically, these two groups were distinct. This seems to eliminate what otherwise might be a tidy explanation: that people engage in more homosexual behavior as their genetic tendency to do so goes up.

So, while the full spread of human sexual behavior may appear to be a spectrum, it seems to be a spectrum that’s pieced together from a number of distinct underlying phenomena. Or, as the authors put it, “These findings suggest that the same-sex sexual behavior variable and the proportion of same-sex partners among nonheterosexuals capture aspects of sexuality that are distinct on the genetic level, which in turn suggests that there is no single continuum from opposite-sex to same-sex sexual behavior.” That’s significant, because a number of psychological profiling methods treat things as a spectrum, including the Kinsey scale.

In some ways, it’s easy to view these results as showing that homosexuality is not really much different from height: “influenced by the small, additive effects of very many genetic variants,” in the authors’ telling. But the results make a strong case for a larger conclusion: sexuality is much more complex than any of our simple ideas might have suggested. Not only is there no “gay gene,” but males and females have genetic influences that only partly overlap. The same is true for people who occasionally pair up with same-sex partners and the exclusively homosexual.

The results also make it clear that homosexual preferences are deeply tied in to other fundamental personality traits, like being open to new experiences. And, while environmental influences may help determine how preferences get formed, it’s clear that social influences can dictate how those preferences get expressed.

Recognizing all this staggering complexity, the authors advise that their results be treated with caution. “Our findings provide insights into the biological underpinnings of same-sex sexual behavior but also underscore the importance of resisting simplistic conclusions,” they write, “because the behavioral phenotypes are complex, because our genetic insights are rudimentary, and because there is a long history of misusing genetic results for social purposes.”

Science, 2019. DOI: 10.1126/science.aat7693  (About DOIs).

https://arstechnica.com/?p=1560631