When Shelley Salling, a pharmaceutical sales representative in Kennesaw, Georgia, took a 23andMe genetic test, she expected to learn a lot about her ancestry and her health. But her "trait report" also told her she was likely to prefer chocolate ice cream to vanilla, to be afraid of public speaking and to get up late in the morning.
So, what's going on here? Could these personality traits really be embedded in your DNA?
Advertisement
Most services like 23andMe, Ancestry and The Geographic Project started off as tools to help people uncover where their ancestors came from. Public enthusiasm over that capability soon led to expanded DNA test options that looked at health predispositions and explored traits that have to do with taste, smell, abilities, physical appearance and more. Some of the answers are fairly easy to reconcile, as most people already know whether or not they have blue eyes, dimples, a widow's peak or red hair. Others, however, such as ice cream flavor preference, fear of public speaking and hating cilantro might be hard to fathom as DNA-driven.
Turns out it's not as simple as one isolated public speaking or ice cream flavor gene. 23andMe enlisted thousands of participants to answer a whole bunch of questions and submit their DNA. Then, it looks at the participants' answers and their DNA to establish connections. If a lot of people who prefer chocolate over vanilla share similar genetic markers, then those markers are identified as possible ice cream flavor predictors. The same goes for fear of public speaking, asparagus odor detection, musical pitch and so on. When a new person's test comes in, the company checks the DNA against the existing genetic data but that's not where it ends. "Based on your genetics, and sometimes your age, sex, and/or ethnicity, we create your personal Trait predictions," says the 23andMe site.
So, it's not just hard-and-fast DNA evidence in play here. "What 23andMe does is they also look at people's sex and how old they are and create a statistical model and make a prediction," explains Carl Zimmer, author of "She Has Her Mother's Laugh: The Powers, Perversions and Potential of Heredity." "When they make a prediction it's better than just a random flip of the coin. It's around 60 or 70 percent, which is pretty good." However, he notes, "That alone doesn't tell you it's all genetic. The marker might not have anything to do with it. It could just be a statistical fluke."
So, some of the results could be accurate and others way off. In Salling's case, she was told she likely had detached earlobes, was more likely to have bunions and less likely to have a cleft chin In reality, she doesn't have detached earlobes, has never had a bunion and does have a cleft chin. "I'm also a dedicated early riser, but the test said that I likely wake up at 8:09 a.m., which is definitely off," she says in an email.
Advertisement