We Don’t Mate Randomly & that Screws up Genetics Research!

This post contains a video, which you can also view here. To support more videos like this, head to patreon.com/rebecca!

Transcript:

Many years ago now, I got a free 23 and Me genetics testing kit. I even made a video about it, and about how I don’t necessarily recommend everyone do it (due to privacy concerns and also general fear-mongering) but personally I was curious to try it and see what there is to see in my DNA. Because as geneticists examine more and more data, they’ve found more and more genes that correlate to more and more human traits. So, I can browse through and see that, for instance, my genes show that I am more likely to have detached earlobes (which I do), more likely to have wet earwax (which, unfortunately, I do), more likely to have blue eyes (yep!), and even more likely to have a unibrow, which technically I did before I was about 14 and my cousin strapped me down and ripped it off so forcefully it was too embarrassed to grow back.

That’s pretty impressive but it was a bit cherry picked, because I DO have thick hair, I don’t hate chewing sounds, I don’t get bitten by mosquitos very often, and I do sneeze in the sun despite my genes apparently making me less likely to have those traits. There are a number of possible reasons for those misses: for a start, genetics are not fate. It’s not nature VERSUS nurture, but “AND.” Our genes and our environment both impact us. And perhaps an even bigger problem is that we simply do not know enough about genetics yet, and by “we” I mean our society of laypeople but ALSO experts. Take misophonia, for instance, which is the extreme dislike of people noisily chewing. I don’t LOVE to hear other people eating but it doesn’t disrupt my day, so why does 23 and Me think it likely does?

Well, that’s because 23 and Me did their own survey of 80,000 people with European ancestry and found that 20% of them hated the sound of people chewing, and those people tended to share a genetic marker “located near the TENM2 gene, which is involved in brain development.”

You watch this channel so you probably already know that correlation does not equal causation. Geneticists use a large amount of data spread across the population to rule out mere coincidences: it’s probably not random chance that 160,000 people tended to have a particular genetic marker and a weirdly intense hatred of chewing sounds. But we also don’t know if there’s not some other factor at play. And in fact, we don’t even know if there’s still absolutely no genetic basis the hatred of chewing sounds at all.

Why? Well, because when identifying these kinds of correlations, geneticists tend to start with one big assumption: people choose mates at random.

“Hold on,” you may be thinking, “are geneticists as a population so socially inept that they think people tend to flip a coin when deciding whether or not to date, fornicate, or more accurately have offspring with a potential partner? Do geneticists think all parties are key parties where you go home with whoever has the 1987 Chevy Cavalier?

No. This is a known issue best summed up in a joke: a farmer was having trouble getting his cows to produce milk and after he ran out of ideas, he turned to the academics at the local university. A multidisciplinary team was assembled, led by a theoretical physicist who was sure they could find a solution. After many weeks of intense work, the physicist went to the farmer in triumph. “I have the solution,” he announced proudly, “but it only works on a spherical cow flying through a vacuum.”

The spherical cow illustrates the problem researchers across many fields have in simplifying complicated real-world problems enough that they can better understand them but not so much that their final conclusions are completely useless. In the case of population genetics, mate selection is an overwhelmingly complicated variable: preferences differ across cultures, and people might choose mates based on physical traits, social skills, finances, education, sense of humor, or according to my aunt the ability to make a really great pie. “Oh, you don’t want to learn your grandmother’s apple cobbler recipe? Good luck ever finding someone to love you.”

The issue is so complex that to do anything in the field without spending ridiculous resources, you pretty much have to ignore it and hope that if you take enough data from enough places, you might smooth out any variation caused by humans and their pesky “preferences” for who they have kids with.

All of which brings me to this latest study, which wondered exactly how much mate preference actually affects the genetics research that relies on the assumption that it doesn’t affect much. You may get an idea of their findings in the title, published this month in Science: Cross-trait assortative mating is widespread and inflates genetic correlation estimates. There’s an open access preprint of the paper available but let me warn you that if you aren’t a geneticist it’s pretty dense, so instead I recommend this article written by the lead authors for the Conversation. Here’s the basic premise: let’s say there are a bunch of dinosaurs and they have either short or long horns, and either scaly or smooth backs, both traits that are controlled by genetics. find a large population of dinosaurs that all share the genes for both long horns and scaly backs. So, we assume that those genes are somehow related to each other, and over on Dinosaur 23 & Me we learn things like “dinosaurs with long horns are also more likely to have scaly backs.” But that’s only if we assume that dinosaurs mate indiscriminately – if we go out to dinosaur nightclubs or dig into the data from Dino Tinder, we learn that for some reason dinosaurs who have long horns prefer to mate with dinosaurs with scaly backs, and vice versa, meaning that their offspring is likely to have long horns and scaly backs even though genetically speaking those two things have nothing to do with one another.

That’s obviously hypothetical because dinosaurs do not have Tinder, but we can imagine the same sort of thing happening in humans: as one author pointed out on Twitter, maybe men with college degrees are more likely to prefer tall women, at which point “genetic variants affecting height become correlated with genetic variants affecting education.” Or maybe redheaded women tend to prefer sad emo boys. Or maybe people with autism are more likely to prefer to be in relationships with other neurodivergent people.

So this study was specifically looking at how much genetic correlation can be explained by this otherwise ignored cross-trait assortative mating, and their conclusion is “a whole helluva lot.” Some studies that previously found correlation between genes can be completely explained by assortative mating alone. The authors write:

“In our study we examined cross-trait assortative mating, whereby people with one trait (for example, being tall) tend to mate with people with a completely different trait (for example, being wealthy). From our database of 413,980 mate pairs in the U.K. and Denmark, we found evidence of cross-trait assortative mating for many traits – for instance, an individual’s time spent in formal schooling was correlated not only with their mate’s educational attainment, but also with many other characteristics, including height, smoking behaviors and risk for different diseases.

“We found that taking into consideration the similarities across mates could strongly predict which traits would be considered genetically linked. In other words, just based on how many characteristics a pair of mates shared, we could identify around 75% of the presumed genetic links between these traits – all without sampling any DNA.”

They’re not completely X-ing out the past few decades of research linking certain traits: they point out that some trait pairs have enough evidence for correlation that it can’t be explained by assortative mating, like how depression and anxiety go hand-in-hand, as do bipolar disorder and schizophrenia.

But their point is that this mostly-ignored factor needs to be considered going forward as we improve our ability to examine the human genome and its effect on our health. It might not matter much to me whether my eye color and ear lobe shape are genetically linked, but it does matter to the future of medicine, as researchers discover new ways to treat diseases that are linked to our genes. And it’s also worth noting the next time you hear news about something like “if your index finger is longer than your ring finger you’re more likely to get skin cancer,” or whatever. It’s complicated, so wear your sunscreen but don’t panic about it.