The Weird Way That Learning About Your Genes Can Change You

Support more videos like this at!


I’ve made a few videos about genetic testing, most of which involve me sort of defending it being open to the general public. To sum up, I think that companies like 23 and Me are generally good in that it allows nerds like me to learn more about our genetic data, and about genetics in general. But we need to be really careful in a number of respects: first, we have to be careful about what these companies are doing with our data, especially here in the US where we have a truly predatory pharmaceutical industry combined with a privatized healthcare system in which companies are just begging for reasons to deny people coverage, like “pre-existing conditions” that they might claim existed in your genes when you were born.

The other way we need to be careful is in how these companies inform people about their genetic information. Statistics can be really tricky, and it can be hard for people to understand how their genes might increase their risk of certain diseases. Like, what does it mean for you if you have a gene that, say, raises your risk of contracting Alzheimer’s by 50%? Obviously that’s not good, but is it something to lose sleep over? What’s your baseline risk? What are your lifestyle factors that might play in?

You don’t want to spring that kind of information on people without putting it into the right context, because you don’t want people freaking out or even doing something drastic, like getting a double mastectomy as soon as they learn they carry the BRCA1 gene mutation that increases risk of breast cancer.

Now there’s new research published in Nature Human Behavior that suggests a related danger we might need to think more about, and which has as much chance of being beneficial to people as detrimental. Psychologists at Stanford found that telling people about their genes was enough to activate what we colloquially know as the placebo effect. Basically, genes became a self-fulfilling prophecy.

They first had two groups of 100 subjects each run on a treadmill until they couldn’t anymore. They then waited a week and brought them back and told one group that they possessed a gene mutation that made it more likely that they would get winded during physical exercise, and that they would be more likely to overheat. They told the other group that they had the protective form of the gene where they wouldn’t get easily winded. Then they had all of them run on the treadmill again.

The groupings were random, and while this gene does exist, the people in each group didn’t actually necessarily have the mutation. Despite that, the group that was told they would be more likely to be winded couldn’t run nearly as long as the people in the other group. Just telling them their genes made them likely to get winded, got winded.

They followed this up by repeating the experiment but with a gene found to make people feel fuller faster, protecting them against obesity. The people who were told they had that gene mutation ate less in a meal compared to people who were told they did not have that mutation.

It’s a pretty impressive effect, considering that it didn’t even require giving anyone a sugar pill, and it really underlines how important it is for genetics companies to frame their results if it’s that easy for people to take their genetics as destiny.

Just so it’s clear, this isn’t a finding that shows that learning you have genes associated with cancer will lead you to developing cancer as a placebo effect. But the researchers point out that if it’s a gene for lung cancer that comes along with reduced respiration ability, you might find it hard to catch your breath.

Of course, even if it does that we don’t know how long it would last. This study, for purely ethical concerns, had to be very short-term — the subjects thought they had the gene for less than a day before the researchers sat down with them and explained that it was all a big fat lie. Because really, you can’t just sent people out into the world believing they have certain genetic risk factors that they don’t have.

The tricky thing with placebos is that sometimes they work even if you know they’re placebos, so there’s no easy solution to this. All we can do is keep studying it and hope that companies like 23 and Me take extra precautions to make sure they’re not sending people out into the world believing things that aren’t true about their genetic destiny.

Rebecca Watson

Rebecca is a writer, speaker, YouTube personality, and unrepentant science nerd. In addition to founding and continuing to run Skepchick, she hosts Quiz-o-Tron, a monthly science-themed quiz show and podcast that pits comedians against nerds. There is an asteroid named in her honor. Twitter @rebeccawatson Mastodon Instagram @actuallyrebeccawatson TikTok @actuallyrebeccawatson YouTube @rebeccawatson BlueSky

Related Articles


  1. Yeah, you’d be surprised at the defense of genetic determinism, even from people who really should know better. (And yes, if you say “a gene for X”, I’m going to demand a protein for X as well, not just a weak correlation.)

  2. Psychological effects are likely major mechanisms for effects that are considered to be ‘genetic’ via GWAS; very likely mediated through the effects of ethnicity on social status, discrimination and then correlate via SNPs of ethnicity.

    The major statistical method used in GWAS is what is called the ACE model. That model imputes that phenotype is a purely linear and purely additive function of SNPs, environment, and noise. The model specifically imputes that variance due to interactions between different SNPs and environment; SNPi X SNPj and SNP X environment effects are exactly zero. We know that they are not zero, we know there are very strong gene X gene effects (transcription factors are products of genes, transcription *has to* be sensitive to transcription factors). Ignoring this real source of variance and imputing that all variance is purely (and linearly) associated with SNPs generates what is called ‘phantom heritability’.

    If you are a member of an ethnicity which is ‘believed’ to be inferior, why wouldn’t that have the ‘placebo effects’ as observed in the OP study?

    If ‘traits’ of different ethnic groups ‘correlate’ with those ethnic groups because of the ‘conventional thinking’ of whole populations, why wouldn’t people with those ethnicity have ‘traits’ that social peer pressure imputes people of that ethnicity to have? GWAS would then be able to associate SNPs of the ethnicity with the trait; even though the trait is purely a social phenomena and has nothing to do with any genetic causation. The causation is by the social environment.

    I am pretty sure that virtually all of the ‘findings’ of GWAS are due to this effect; they are due to the flaws of the ACE model, not due to any genetic ‘causation’ of the trait. If a trait is ‘important’, then it is ‘too important’ to be ‘set’ by genetics. It has to be set by development, so that the organism can adopt how much, or how little of the trait is necessary, based on what would be better, in the particular environment the organism finds itself.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top button