The Epigenetics of Sexuality - Wrong on So Many Levels

The ludicrous amount of coverage of a basically uninterpretable experiment suggests a surprising amount of interest in the basis of sexuality, and I am even more irritated by this than I am by poor use of statistical analyses.
This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.

Another day, another truly execrable epigenetic report. Epigenetics is the study of chemical modifications to our genetic material which influence how genes are expressed. It provides the mechanistic link between our genes and our environment, and is a beautiful area of biology. It is involved in phenomena as diverse as the flowering times of certain plants and the gender of crocodiles, and from novel treatments for cancer to the coat color of calico cats. In the last few years scientists have developed new techniques to analyze the patterns of epigenetic modifications on the genome, and there is a tsunami of papers emerging. And unfortunately some of them are very bad indeed.

The latest awful study hasn't even been published yet, instead it was presented at the annual meeting of the American Society of Human Genetics. Researchers from the University of California Los Angeles took saliva samples from pairs of identical twins, some of whom were gay and some straight. They analyzed the epigenetic modifications and announced they had found five that together would predict the sexuality of the donor accurately 67 percent of the time. The authors also speculated that this may give us insights into the cause of homosexuality.

Does 67 percent of the time sound good to you? Try thinking of it another way. Would you bet anything beyond a dollar on something that will be wrong a third of the time? In yes/no situations, you could flip a coin and be right 50 percent of the time on average. Sixty-seven percent doesn't sound so impressive now, does it?

And that's not the only problem. The number of twin pairs was far too low to generate any genuinely meaningful statistical data. Saliva samples contain a mixture of cell types that may have different epigenetic modifications, and why would saliva be a relevant tissue in which to study sexuality anyway? And there are other technical problems, but I am pretty sure you've recognized my direction of travel by now.

So why am I wasting my time on an unpublished piece of badly performed science? By rights, this should have sunk without trace, not because it's controversial but because it isn't much good. But Nature, the world's leading scientific journal, wrote a piece on it and then the non-scientific press picked this up and gave it a ridiculous amount of coverage. I can't quite bring myself to castigate the popular press, but Nature? This really should be placed in your "What Were We Thinking?" filing cabinet of shame.

The ludicrous amount of coverage of a basically uninterpretable experiment suggests a surprising amount of interest in the basis of sexuality and I am even more irritated by this than I am by poor use of statistical analyses. I think of myself as a naturally curious person, it's why I am a scientist. I wonder about loads of stuff all the time -- how cell phone signals move, why dead dragonflies fade but butterflies don't, why most people are more interested in life on Mars than in their own neighborhood. But with the exception of wondering what the Kardashians are up to, there is nothing that I am less curious about than why I am a lesbian. And I think that is very common. I know few gay people who waste any time on this. In contrast, it is almost always a straight person who asks this question of someone who is gay.

But here's the thing. I have never heard a straight person wonder why they are straight. No matter how it's dressed up, the question is always one of investigating the deviation from the norm. The defense that is usually put forward is that studies such as the recent one are just looking to explore the biological basis behind a range of human behaviors. But that position is both naïve and potentially dangerous.

I am old enough to remember a former chief rabbi in the UK stating, on the premier radio news program, that a test for homosexuality would be a good thing because pregnant women could choose to abort affected fetuses just as they would for any other disease. I can marry my partner now, but when we first got together our relationship was defined in British law as "pretend." Although things have generally improved in the west, would you want to be gay or lesbian in Uganda? South Africa? Yemen? Tennessee?

Science doesn't operate in a neutral vacuum divorced from wider society. Every one of us has a responsibility to think about how our research will be interpreted and used. The very questions that we phrase demonstrate our biases and assumptions. And the odd thing is that whenever we attempt to apply simplistic biological algorithms to complex and charged human conditions, we usually end up with bad science and a lot of hype. Look up any study on race and intelligence and you'll see what I mean. Human cognition and emotions are essentially too complex to be properly modeled by childishly simple algorithms, however comforting those might seem at first glance.

I know some of the responses to this blog will be that I am a Luddite. I will try to respond quickly, but I should warn you that I may be too busy Googling the latest exploits of Kim and Kanye.

Popular in the Community