23andMe effectively shuts down (Dec 7th, 2013)

23andMe Logo

Try typing “23andMe scam” into Google. You’ll be directed to my most popular blog post of 2013, in which I posted the results of my genetic test with 23andMe. It’s the third most popular search term leading into this website (behind Buddhism and Hunger Games in 1st and 2nd place). 🙂

You can read the original post here.

In December 2012, I spat into a tube and sent it to 23andMe, a genetic testing company in California. In January 2013, they sent me the results of my genetic information over several days. I sifted through swathes of data and posted all the conclusions worth hearing onto my blog. There were 12 of them.

As someone trained in biology, particularly in genetics, I knew how to interpret the genetic data. I knew to ignore all of the health information, especially the vague correlations in tiny studies with genes of unknown function, and posted only the inconsequential genes and the ancestral information instead. This wasn’t a privacy concern—I posted all the findings that I found genuinely interesting. But I did omit any findings that I knew to be total bunk.

In one blatant statistical blunder, 23andMe told me I had a 1.5x greater chance than the “average person” of getting cleft lip. In another, they told me I had double the likelihood of getting certain diseases for which environmental factors were by far the greatest predictors. 23andMe doesn’t consider environmental factors and doesn’t differentiate between meaningful and meaningless statistics.

People not trained in genetics or statistics wouldn’t have the same level of insight that I did. Some customers could read all the misleading data like numbers on a die, and then make false conclusions as a result. 23andMe never did enough to protect its customers from such confusion. In fact, they passively encouraged it with gimmicks like “Genetic Melody” and the inclusion of vague correlation studies in tiny populations for very serious diseases. This data could worry people unnecessarily. While 23andMe communicated science much better than did the average newspaper, there was still a gaping linguistic chasm between scientists and consumers. The way that 23andMe reported health information to the public was still in need of major improvement.

But that improvement didn’t come quickly enough, and last week, the FDA shut down 23andMe’s entire genetic health reporting service. Its ancestry service lives on, which has always been 23andMe’s most interesting component in my opinion, and only new customers (after November 22, 2013) will be affected. Nevertheless, this decision is a huge blow for 23andMe and a momentous victory for advocates of science communication.

Rightly so. The fact that so many people are typing (and I see you!) 23andMe scam into Google indicates that a problem is at foot.

So while I’m pleased it’s been shut down, I also regret not making these shortcomings clearer in my 23andMe post one year ago. Despite the comments and emails I received from people trying to make sense of their 23andMe results, I forgot that not everyone is science-literate.

Fortunately, the FDA remembered. Thank you.


1 thought on “23andMe effectively shuts down (Dec 7th, 2013)

  1. >As someone trained in biology, particularly in genetics…
    Can you provide a reference to your publications in this area?


Talk to me

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s