23andMe and Personalized Genetic Testing (CS04)

This week on cracked science: What to do if
you want to spend 250$ To find out if you have back hair. [title] Hey, this is Jonathan Jarry and you’re watching
Cracked Science, the bi-weekly show from the McGill Office for Science and Society that
separates sense from nonsense on the scientific stage. Let’s talk about your genes. No, I don’t mean the 16-hundred-dollar pair
of distressed pants you bought online and seriously hope your significant other won’t
find out about. I mean, of course, the bits along your DNA
that help make you… you. If I asked you to guess who is running the
largest genetic study ever, you may think of a university, maybe Harvard or one of the
universities of California, of which I am reliably informed there are now 57. The actual answer may surprise you. It’s 10. They have 10 campuses. No, but the people running the largest genetic
study ever are 23andme, the company whose motto, “welcome to you”, sounds like the name
of a bad romantic comedy from 2003 starring Matthew McConaughey. 23andme is a company that offers “direct-to-consumer
genetic testing”. This means that you no longer have to go see
a medical doctor to gain access to your genetic information; You can go through a business. Before we talk about what 23andme is now planning
to do with your genetic information-and, yes, I mean you, as you may be one of the companies’
million plus customers-, we need to talk about direct-to-consumer genetic testing, its promises
and its reality. It is an industry that is predicted to grow
to 340 million American dollars in the next five years and is appealing because genetics
is cool (unless portrayed by Hollywood) And because we all feel a sense of entitlement
over our genetic information. So, let’s look at what a customer of 23andme
gets for the price, and keep in mind that there are other companies offering similar
services but, to simplify, let’s look at the most popular one. 23andme will have you spit in a vial, they
will extract your DNA, and they will test it for a long list of variants that can be
found throughout your genome. You will get back a whole slew of reports. Some are about non-health-related traits. For example, and I am not making this up,
they will test to see if you have the DNA variant associated with back hair (Note: Available
for men only). Sorry, ladies. You’re really missing out on that one. Other traits you will be informed about include
whether you have light or dark hair; What your toe length ratio is; And whether or not
you have a unibrow! Keep in mind: You are paying good money for
information that an ancient roman could have figured out using their eyes and a bronze
mirror! But more importantly, you will get reports
on mutations you may carry that could cause a disease in your child-a disease like cystic
fibrosis-and reports on risk factors for things like Alzheimer’s disease. Now, back hair aside, you may believe that
this is valuable information, but many scientists disagree. For example, you may think that the information
contained in your genes is highly predictive, the same way that the phrase “a film by Michael
Bay” is predictive of, and I quote a popular movie critic, an “awful lingerie and beer
advert that will carry on forever”. But the truth is, the idea that every disease
is caused by a mutation in a single gene was very naive when first thought of. Here’s an illustration of how much more complicated
reality is. The small area in green represents the kinds
of diseases that can be easily predicted based on genetics alone, whereas everything else
is in a shade of red, meaning genetic tests do a poor job of predicting what will happen
to you. It’s like an average person betting on horses:
Sure, it’s easy to figure out that this horse would never win the race, but if every other
horse looks like this instead, the average person will have a hard time betting on the
winner in advance. And here’s a table that shows how much influence
genes have on complex diseases, like prostate cancer and type 2 diabetes. These are diseases that are influenced by
more than one gene as well as by your environment. As you can see, apart from a few, none come
close to 100%. Then there’s the problem that different testing
companies will give you different risk profiles based off of the same DNA. The U.S. Government accountability office sent the
DNA of volunteers to four different companies back in 2006 and wrote that these “fictitious
consumers received test results that are misleading and of little or no practical use.” You can see, for example, that one volunteer
was told that his risk of prostate cancer was average by two genetics companies, below
average by the third, and above average by the fourth. Which makes these results as reliable as a
Ouija board! And the people who man the phones at these
companies may not always be statistically literate genetic counsellors. Here’s a phone exchange between an undercover
agent of the government accountability office and one direct-to-consumer genetic testing
company: “so, if I’m high risk, does that mean I’ll
definitely get breast cancer? “you… you’d be in the high risk of, you know, pretty much
getting it.” Oh… my god. This is what happens when DNA companies hire
Miss Cleo to do customer service. Now, the risk profile *I* would get out of
a test like this is already dubious, but if you are African American or Asian, you may
be extra disappointed. You see, to interpret your risks, these companies
need a baseline in your ethnic group, because how common a variant is changes from population
to population. A piece on website Quartz mentioned that 23andme
issued a report to its Korean writer in which the writer’s genetic profile could only be
compared to 76 Koreans, which a geneticist called “a really low number.” Yeah, no kidding. In fact, a number of risk assessments made
by companies like 23andme are based on population studies done on white people. This is not the fault of the businesses themselves;
This is an artifact of the bias of the basic research that is being done. Now, happily enough, 23andme received nearly
2 million dollars from the U.S. Government to fully sequence the genomes of
many of its African American customers, so that this white bias in genetics research
can start to be rectified and we can move away from things like “your profile could
only be compared to 76 Koreans”. So, to summarize, low predictability of the
tests, useless information when it comes to traits, different risk assessments depending
on the company, less information if you’re not white, and Miss Cleo answering your questions. But once you are done worrying about your
slightly increased risk of developing prostate cancer, you can gloss over your custom health
report from 23andme that may state the following: “These habits made the biggest difference
in people with your genetics: Limiting red meat, avoiding fast food, sleeping a healthy
amount, eating vegetables, and exercising.” Thank you, science, for providing me with
highly personalized advice that does not apply to anyone else. If you and four of your friends were thinking
of spending 250$ each on this, you might as well pool your money and buy an ivory butt
scratcher. But even if you don’t give your DNA to 23andme,
it remains the largest genetic study ever, and the company has recently decided to do
something interesting with this information: Drug discovery. No, the company cannot look at your DNA and
know that you like to smoke some ganja while watching Up in Smoke. But they can look at a drug that already exists
and that targets a particular gene, and try to find out, by screening their vast library
of DNA samples from people like you, if some customers have both an interesting variant
in that gene *and* a particular disease, because it could be that that genetic variant causes
the disease and that the drug could help treat this disease. And I think this is a great initiative, though
we shouldn’t have unrealistic expectations, as drug discovery is notoriously difficult. For example, any researcher who is teaming
up with 23andme to look for associations between customer characteristics and their genetic
information will have to rely on self-reports. For example, if you want to look at customers
of 23andme who have clinical depression, you depend on these customers accurately self-reporting
their diagnosis of clinical depression. Some may not be willing to share this information;
Others may feel sad and think that this is clinical depression. This is something that researchers must be
acutely aware of when partnering with 23andme. But it is certainly ironic that, while many
people expect to be paid to participate in research studies, customers of 23andme are
willing to not only give them their DNA, but to pay them as well for the privilege, and
all for the pleasure of finally finding out the answer to the most important question
in the universe: Do I have back hair? Yes, Andy Garcia, yes, you do. That’ll be 250$, please. My recommendation this week is a very comprehensive
piece about 23andme by Erika Check Hayden and published in the journal nature. It’s called “The rise and fall and rise again
of 23andme”. It covers ground I didn’t have time to address,
such as the company’s bizarre response to the FDA and an interesting portrait of the
woman in charge of the company, who seems to have taken inspiration from Mark Zuckerberg
when designing her nameplate. That’s it for this week. I’ll be back two weeks from now with a whole
new show. In the meantime, don’t forget to comment,
like, subscribe, and check out our website at mcgill.ca/oss.


Add a Comment

Your email address will not be published. Required fields are marked *