[Originally published as Large Study Indicates Genetics Has Little Influence on Sexual Orientation]
I don’t normally write on topics like this, because studying human behavior is a tricky subject. There are all sorts of different explanations for a given behavioral characteristic in people, and trying to isolate a specific cause is difficult, to say the least. However, there has been a lot of news about the recent study that concluded there is “no gay gene,” and I have gotten several questions about it. As a result, I decided to read the study and share my thoughts.
First of all, it’s not surprising that there is no gay gene. In fact, researchers have said that for years. If there were a single gene that heavily influenced whether or not a person is homosexual, it would have been easy to find and discovered years ago. Also, even something as simple as the color of your eyes is governed by at least eight different genes. Thus, to think that something as complex as sexual behavior is governed by one gene is naive at best. So that specific result of the study is not even interesting, much less newsworthy.
What makes the study newsworthy is its size, its scope, and the fact that its conclusions are very weak.
The study is massive in two ways. First, the main study looked at 477,522 individuals, but then it repeated the study using three smaller datasets that were composed of 15,142 individuals. Whenever you study people, the more people you have, the less uncertain your results will be. Thus, the sheer number of individuals in the study makes it important. Second, the study compared the entire genomes of the individuals. In other words, they looked at all the DNA found in the nucleus of the individuals’ cells. That’s a massive amount of data for a massive number of people!
What it tried to do is compare single nucleotide polymorphisms (SNPs) among the individuals to see if they could be correlated with sexual behavior. If you aren’t familiar with the term, SNPs are the most common variation between sets of human DNA. Genetic information comes in sequences of DNA building blocks called nucleotide bases. There are roughly three billion nucleotide bases in one strand of human DNA. An SNP is a change in one of those nucleotide bases.
The entire genomes of the individuals studied were available to the researchers, as were answers to several questions about behavior. One of those questions was whether or not the individual had ever had sex with another member of the same gender. When they tried to find SNPs associated with people who answered “yes” to that question, they found five, none of which were on the sex chromosomes. When they replicated the study with the three smaller samples of people, they could only confirm three of those five.
However, all of these SNPs were very, very weakly correlated with the individual’s answer to the question. One of them, for example, made a male 0.4% (four tenths of one percent) more likely to have answered “yes” to the question. That’s an incredibly small effect. Given the weak effect of those SNPs, the researchers concluded that there must be many, many genes associated with sexual behavior, which is not surprising.
As a result, they decided to just look at the similarity in the genomes of the people who answered “yes” to the question. This is called a “heritability” study, and it is typically used to estimate the contribution that DNA makes to a specific trait. When they did that, they found that genetics can only account for 8-25% of whether or not a person answered “yes” to the question. This is significant, since behaviors that have been linked to genetics typically have much higher heritability results. For example, current studies on alcoholism indicate that 40-60% of whether or not a person is alcoholic is determined by genetics. A score of 8-25% indicates that genetics has a tiny effect.
There were two other interesting aspects of the study. First, they looked at the genetic similarities among the people who answered “yes” to the question and tried to correlate them to other traits like alcohol use, anorexia, autism, loneliness, etc. They couldn’t find any genetic correlation with most of the traits, but they did see a genetic correlation with risky behavior, smoking pot, schizophrenia, ADHD, loneliness, and having a large number of sexual partners.
This highlights the main weakness of the study’s focus. Answering “yes” to the question of whether or not you ever engaged in sex with the same gender doesn’t really probe your sexual orientation so much as your willingness to try new things and the culture with which you are engaged. I would think it is rather obvious that if you get high a lot, you are more likely to try a homosexual experience than if you never get high. In the same way, if you have a lot of sexual partners, it makes sense that you might decide to try at least one of the same gender. That’s what the correlation seems to be saying.
This brings us to the second interesting aspect of the study. The authors tried to address this weakness, but it was difficult. For most of the individuals, they had data related to how common homosexual experiences were for them. Among those individuals, there was a wide range of variation from mostly engaging in heterosexual experiences (but making rare exceptions) all the way to being exclusively homosexual.
If they looked at the genetic similarities they had already found and tried to correlate them with the percentage of homosexual encounters, they found weak correlations that had an enormous amount of error and were different between the sexes. The correlation was 0.72 +/- 0.55 for men and 0.52 +/- 0.68 for women. A “0” would mean no correlation, and a “1” would mean perfect correlation. The “+/-” indicates the error involved. In other words, the error involved is almost a much as the correlation found for men and more than the correlation found for women. Once again, this indicates that if an effect is there, it is very, very weak.
If this study’s findings are confirmed by other studies, it means that genetics plays a very small role in determining a person’s sexual orientation. This is important, since the same kinds of studies indicate that genetics plays a much more significant factor in other behavioral traits, such as alcoholism and ADHD.