Why I didn’t do 23andMe

I had a free voucher from the summer, but decided not to go ahead with a saliva test.  Here are my reasons:

1. I thought there is option value and I can always do a test later, for a better and more accurate service.  (I hadn’t thought of the FDA shutting the whole thing down, but still I expect the service will return in some manner, if only under another corporate banner or from overseas.)

2. I thought the “worry cost” of negative information would exceed the benefit of whatever specific preventive measures I might take.  Most useful ex ante preventive measures, such as diet and exercise, are fairly general in their application and I didn’t think there was likely much to be learned about specific measures for specific potential maladies.  And here is an interesting short piece on the likelihood of false negatives.

3. One might take more preventive measures with one’s ex ante and more uncertain knowledge than with one’s ex post and more certain knowledge.  For instance an absence of negative information might have encouraged me to slack on exercise, to the detriment of my eventual health outcomes.

4. I wouldn’t describe privacy concerns as my major worry, but at the margin still they counted for something.  I felt eventually this service would prove equivalent to making my genome public information, via something called GenomeLeaks or the like.  Why do that without having a better sense of its longer-run implications?

I’m glad I didn’t do it, glad I had the choice to decline to do it, and I am still feeling no temptation to do it in the future.  I do feel a slight amount of guilt for not contributing to a future “Big Data” project, but so be it.  I also am glad I am not contributing to some of the inevitably unethical uses to which eugenics will be put, and that is more than a counterbalance, given that I expect no practical benefit from reading my own test results.


I was only interested in the ancestry information. I wish they offered a test that only provided that information. On a related note, this is amazing: http://scienceblogs.com/gnxp/2008/08/31/genetic-map-of-europe-genes-va/

Anne, Ancestry.com apparently offers a DNA-analysis service that focuses only on ancestry. Probably ties into the rest of their geneological databases as well, I imagine. http://dna.ancestry.com

I looked at http://www.familytreedna.com after someone mentioned it here. I'm male, so I could potentially buy any combination of "Family Finder", "mtFullSequence" and either Y-DNA37, Y-DNA67 or Y-DNA111. All three would be $457, and it's not clear that you get any more geneological information than what 23andme provided.

buddyglass - I've tested at familytreedna, 23andMe, and ancestry.com. If you did do all of those tests, you would indeed have more genealogical information, as you would get matches on both your patrilineal (Y-DNA) line and matrilineal (mtDNA) lines which 23andMe doesn't give you - it only estimates what your haplogroup designation is. So, for example, you and I might share the same haplogroup, and either FTDNA or 23andMe would give us that info. But that doesn't mean we're related in "genealogical time". FTDNA would tell us if we matched on our Y chromosome, 23andMe cannot. It's a totally different test. The really expensive test at FTDNA is the mitochondrial one, and that is certainly the least useful one for genealogical purposes. As a male, a combination of Family Finder (comparable to, and same price as 23andMe's test) + Y-DNA test would give you a lot of information to work with. If you're interested you can read more about all these tests on this blog: www.yourgeneticgenealogist.com - or on any other of various sites you can find by Googling...

Thanks for the comment and info. I realized after the fact that familytreedna's "Family Finder" is the same sort of autosomal test as 23andme's. What confused me is that I knew 23andme gave haplogroup information and I thought that only came from doing Y and mtDNA tests. I'm mostly interested in the "large scale" genealogical info as opposed to "cousin finding", so the autosomal test is all I really need. Given that, which of the $99 tests would you recommend: 23andme, ancestry.com or FTDNA? Once upon a time I had an account at ancestry.com, but I shut it down after doing a bunch of research and exporting it to a GEDCOM. The monthly fee was killing me.

Yes, amazing, but it's somewhat contradictory, since like the article says the genetic diversity of Europeans is rather low, say compared to Africans (where it is huge I've read). But this part is interesting: 'Both teams found that southern Europeans boast more overall genetic diversity than Scandinavians, British and Irish.'. Yep, just like my grandpa says, the Greeks are superior to every other race. Yep yep yep.

With their data they can say more about the diversity of Europeans than the diversity of Danes. ("Do note the small sample sizes for some groups, the N for Denmark, Finland, Latvia, Slovakia and Ukraine is 1.")

(While both Danes and Greeks messed about in boats, the Danes mostly on the northwestern fringe, the Greeks at the crossroads.)

Ray, giving how the regions vs genetic diversity are doing, grampa may need to re-assess.

Either way, I just call us thoroughbreds.

It would take many generations of Tyler's "inevitably unethical uses of eugenics" for anyone to become a thoroughbred.

You could use the 23&Me service while declining to participate in their big data project. I believe it kills their revenue model, but they give you the option (which they probably hope few people take).

[ "... but still I expect the service will return in some manner.." ]

Rest easy. DNA collection/analysis/archiving will be mandatory for all by year 2025, under ObamaCare III.

It will also mesh very well with the DHS/FBI/NSA dossiers on all Americans. Fingerprints and photographs are so 20th Century.

'DNA collection/analysis/archiving will be mandatory for' anyone the police arrest and want to test, as of 2013.

'The Supreme Court has ruled criminal suspects can be subjected to a police DNA test after arrest -- before trial and conviction -- a privacy-versus-public-safety dispute that could have wide-reaching implications in the rapidly evolving technology surrounding criminal procedure.

At issue in the ruling Monday was whether taking genetic samples from someone held without a warrant in criminal custody for "a serious offense" is an unconstitutional "search."' http://edition.cnn.com/2013/06/03/justice/supreme-court-dna-tests/

No need to make fantasies of something that won't happen, compared to the mechanisms already in place today. And do note, those governmental mechanisms have absolutely nothing to do with the provision of health care.

Sort of like how fingerprinting was first a feature that young drivers were subject to, but which is now also becoming routine for all.

Sounds closer to the Republican-led Patriot Act to me, citizen.

Republican-led Patriot Act

Let me fix that for you:

political-class-led Patriot Act

Anyone who believes only Ds or only Rs have all the answers or are evil and not to be trusted is hopelessly naive. Acton's warning still applies:

"Power tends to corrupt, and absolute power corrupts absolutely. Great men are almost always bad men."


When a person gains power over other persons–political power to force other persons to do his bidding when they do not believe it right to do so–it seems inevitable that a moral weakness develops in the person who exercises that power. It may take time for this weakness to become visible. In fact, its full extent is frequently left to the historians to record, but we eventually learn of it. It was Lord Acton, the British historian, who said: “All power tends to corrupt; absolute power corrupts absolutely.”

Please do not misunderstand me. These persons who are corrupted by the process of ruling over their fellow men are not innately evil. They begin as honest men. Their motives for wanting to direct the actions of others may be purely patriotic and altruistic. Indeed, they may wish only “to do good for the people.” But, apparently, the only way they can think of to do this “good” is to impose more restrictive laws.

“Power tends to corrupt, and absolute power corrupts absolutely. Great men are almost always bad men.”

Therefore, God is the Devil and the Devil is God.

I didn't see any follow up news items after the FDA sent them the C&D letter. Are they going to bring themselves into line with regulatory norms, and continue operating?

The CEO sent out a letter to customers Wednesday stating "I apologize for the limited response to the questions many of you have raised regarding the letter and its implications for the service. We don't have the answers to all of those questions yet, but as we learn more we will update you." So no answers yet AFAIK.

Apparently they were given 15 days from the date of the (November 22nd) letter to explain how they will make themselves compliant. I imagine after the storm of negative press they received, they'll make more of an effort to operate above the table this time.

#2 is weak sauce. Disturbingly unscientific mindset. Are you saying you are a prisoner of cognitive biases that will interfere with your ability to process information rationally? Maybe, but no way to go through life.

It's not weak sauce, 'worry costs' can alter behavior in unfortunate ways. Two years ago I acquired a medical diagnosis and doctor, who regularly lectures me that 'people like me' are not physically capable of the career path I am on. At least statistically she may have a point and I do try harder to respect my limits. However, I am very happy she was not 'coaching me' when I chose that career path. Even if I step off and do something more suited to that part of me, I would not have wanted to 'miss the dance.' More information (even if it's the truth) is not always better.

You strike me as a tough old coot Claudia. Ouch! You strike me! But seriously, I've read in Japan if an old person has terminal cancer the Japanese doctor is ethically obligated to lie to them, to give the patient peace of mind. Maybe the doc would however tell the next of kin to gently remind the patient to get their affairs in order, such as a last will and testament. And maybe do all those1001 things they wanted to do before they die.

Ouch! I don't thinks she's all that old, Ray.

Well yeah, TMI. But that generally refers to useless, extraneous information. It sounds like you're saying there is some tendency or proclivity about yourself that you would prefer to be unaware of.

Know thyself, I say.

Know thyself sure, but don't expect 23andMe or anyone else to tell you who you are, I say. Self aware does not require fully informed.

PS TMC, I suspect 37 is ancient for Ray's women, no offense taken, made me smile.

“The fundamental cause of the trouble is that in the modern world the stupid are cocksure while the intelligent are full of doubt.”

― Bertrand Russell

Or Yeats's version, "The best lack all conviction, while the worst are full of passionate intensity."

Or, Charles Bukowski: “The problem with the world is that the intelligent people are full of doubts, while the stupid ones are full of confidence.”

catchy aphorisms. but untrue. try arguing with a new yorker reader about sex differences. or yglesias about the nature of his polar bear hunting attackers. full of doubts, they are not. or at least they don't let on like they have doubts.

I do not feel the need to use this service, but any time that any arm of the government tries to legislate my rights over my own body, and without a damned good reason for it, well let's just say that it both pisses me off and makes me lose faith in government.

Whose interests are protected by outlawing a test like this? Who are the potential victims? Who stands to gain? Why does the FDA think that I should have less, rather than more, information about my very own body.

My body, my life, my decision. The FDA can go to hell. I'm not American.

Nathan, There are any number of ways to get your genes sequenced. The issue is their testing and the content of the information they provided consumers.

One way to get them sequenced for free is to commit a crime.

I expect DNA samples taken by the criminal justice system are not sequenced but used to generate a "DNA fingerprint" as it would be a much simpler task sufficient for matching a person to DNA "beyond a reasonable doubt"

'I thought the “worry cost” of negative information would exceed the benefit of whatever specific preventive measures I might take.'

Upon what possible basis? This doesn't make much sense. Why shouldn't you have worry costs now? Just because your psychological temperament means you don't have worry costs now doesn't mean you have nothing to worry about. You should have the meta-worry that you're not worrying enough - that your default state of ignorance-bliss is leading you away from easy fixes that yield an even better equilbrium ('knowledge-bliss'?)

Think of it this way, is it useful for a woman to know she's pregnant when she's deciding whether to cut back on the alcohol / cigarettes / drugs, etc? Without testing, a woman could be worried / cautious / trading-off-pleasures all the time. With tests, she has an easy way to know when and how to adjust her behaviors.

Some of the risk factors are 'actionable' or easily mitigatable. For instance, I know a sun-loving woman who had a dark complexion (but from a family with most members having fairly fair skin) who nevertheless discovered she was at particular risk for several disorders aggravated by exposure to ultraviolet light. If I recall correctly, I think among them was a form of skin cancer and macular degeneration.

East prescription: 'Stop sunbathing'. People from fair-skinned heritages who sun-bathe could just as easily have worry as their default presumption, take a cheap and easy genetic test, and then be relieved of their worry from the results.

Everyone seems to be so absolutely certain that "more exercise" is good for you that I smell a rat. What's the evidence?

I think tests on rats ( or gerbils?) have proved that more exercise is good for the researchers.More scientific papers. ( jk.)


Obviously "more" depends on where you start. I hike a medium amount and run a little. I feel great when I do. It is surprisingly common though for runners and hikers to become addicted and do themselves damage.

Speaking of people who go too far, the extreme trends of "paleo" and "minimalist running" go too far. But I think their starting point was not too far wrong. We should eat a natural, omnivorous, diet. And we probably all could run a little.

What is extreme about paleo and minimalist running?

When people "do paleo" now, they'll often try to completely eliminate grains. They may try a meat and vegetable diet. This even if "their people" have been farming grains for 5,000 years. Even stranger, there are now industrial, processed, "paleo" foods. Apparently they'll grind down a cow, process it chemically, and return the things that our ancestors would have gotten by chewing on a deer.

The minimalist runners will look at native people who run in sandals and then say "barefoot is best," ignoring what the runners they admire actually do. Sure, some of us go with a good basic and light shoe, somewhat like those running sandals, but many will wear funny toed things that try to look like bare feet. That the funny looking things don't protect like a sneaker, and can lead to bone bruises when trail running becomes apparent later.

One of the weirdest things though is the association between minimalist running and veganism. We are told that native runners ate chia and that gave them all they needed, so that when they ran for hunting purposes ... wait a minute!

You get it by now. Paleo and minimalist running show that anything humans do, they can go nuts on.

There's a fair amount of science discrediting the barefoot running stuff. I recall reading a comprehensive report on the topic and it was hard to not think the barefoot running thing was a way to sell huaraches. I run three miles each morning and I have stepped on enough sharp objects to know why man invented shoes before the wheel.

The paleo diet strikes me as a repackaging of Atkins which is a repackaging of the Stillman Diet. Low carbohydrate diets go back a long way. IIRC, the Salisbury Steak was invented as part of one such diet in the 19th century. I eat few carbs, except at holidays. My diet is meat, eggs, fish and some green vegetables. The reason is carbs make me fat so I avoid them.

Not only that, in much of the world true barefoot running invites infection or parasites. Keep them critters away.

"Ignorance is bliss," Big Brother

'The phrase "Ignorance is bliss" is from the poem Ode on A Distant Prospect of Eton College by the English poet Thomas Gray. The poem was published in 1742. The full context of the quote reads as follows: "Thought would destroy their paradise. No more. Where ignorance is bliss, 'tis folly to be wise". To read the entire poem, go to http://www.thomasgray.org/cgi-bin/display.cgi?text=odec' - http://www.reference.com/motif/society/origin-of-ignorance-is-bliss

And a bit of condensed analysis - 'This is the origin of the now proverbial phrase “ignorance is bliss.” In the poem, it referred to young people who are happily oblivious to the difficulties they will face as adults — and to the ultimate, inescapable fate of death.' http://www.quotecounterquote.com/2010/08/ignorance-is-bliss-except-when-its-not.html

Thank you for making me less blissful!

Option value: Yes, they were about to release an expanded 4th version of their test.

It's not too late to cash in the voucher. Do it before the option value goes away.

Your points 2 and 3 seem a little contradictory. You expect bad news to make you worry, but you expect good news to make you slack off. Can you manipulate your biases to make the bad new motivate?

But, yes, there is almost no informational content in 23andme. Knowing that you have 2x, let alone 10% lower or higher chance of any given disease is useless. The only useful information is if a rare disease is promoted to a common disease, but virtually none of their tests tell you that.

Something stinks about this whole scenario. It seems more and more like an organized hit on 23andme.

Part of it is that there's also a ton of misinformation floating around. Just about every news article on the subject and most blog posts (including this one) claim that the FDA "shut them down". That's not what happened. The FDA only halted marketing of their genetic testing product to new customers. 23andme is still servicing their existing customers. There's also a lot of misinformation that 23andme was promising the ability to test for diseases. That's not true - what they offer is a test for genetic markers that might indicate a higher or lower risk of you getting a particular disease compared to the general population.

The other part is what seemed like a volley of opinion pieces applauding the FDA's decision and decrying the practices of 23andme, many of which appeared within days of the FDA's decision. A small sample:

The main themes in all these hit pieces are fear of big data, fear of disruptive innovation in medicine, and fear of what individuals might do with more information.

I think it is interesting that how we frame issues controls the discussion.

For example, on this site it is framed a freedom and government oppression or over regulation.

But if you went to a medical ethics website, or even a med school ethics class, how would this be framed or discussed.

To get a flavor and to get a broader sense of the issues, google "ethics and genetic testing" to read some of the scholarly articles in the medical ethics field, if you want to expand, rather than narrow, your frame.

It never even occurred to me until now that 23andme was fundamentally another big data project. I guess it's never straight commerce any more, the plan is always to let the real customer screw the "customer."

To each their own! We took the tests because it sounded fun and because I've always enjoyed maps of ancestral genetic traits. No other reasons. I could care less about the health aspects. Knowing or not knowing.... it doesn't affect me because the uncertainty in most of this stuff is so large as to be almost irrelevant. In my experience the concerns of #2 are not valid but I'm sure that varies from person to person. I should suggest that as a test on 23andme! Anyway, we get our results back and find out my wife is a carrier for CF. This stuff is good knowledge to have. Not bad.

Look out below!


Sometimes Tyler gives the impression that he's satisfied sitting poolside as civilization declines.

Notice he doesn't even give a real answer as to whether he agrees with the FDA's decision to regulate direct-to-consumer genetic testing.

You have influence - use it to support or oppose policies more relevant than the antiquated crusade against eugenics.

What do you mean, impression?

Here is a text from an interview with Die Welt, that he himself linked to here ( http://marginalrevolution.com/marginalrevolution/2013/09/assorted-links-905.html ) -

'Die Welt: Stellen Sie sich vor, wir würden bereits das Jahr 2033 schreiben – würden Sie dann noch gern in den USA leben wollen?

Tyler Cowen: Oh ja. Ich bin sehr zuversichtlich, dass ich dann nach wie vor ein angenehmes Leben haben werde. Aber sehr viele andere Menschen werden in 20 Jahren finanziell deutlich schlechter gestellt sein als heute.' http://www.welt.de/wirtschaft/article119754881/China-wird-die-Supermacht-USA-nicht-ueberholen.html

(My translation back into English from what was undoubtedly an interview conducted in English, Which means this translation is definitely not precisely what was said -

Die Welt: Imagine we were writing in 2033 – would you still want to live in the USA?

Tyler Cowen: Yes. I am confident that I would be living a comfortable life, as before. But many other people will be in clearly worse financial shape in 20 years than currently.)

At the point of writing this comment no one else has disputed #1. Also I am a 23andme customer.

While it is possible that the results of the test become more accurate (which I should say 23 gives updated results as they can) without some new information about DNA that is outside of the raw data, there is very little difference between now and 10 years from now. It is reading the same data. I can download my raw data from 23. If another analysis service were to show up, I hope they would allow me to import my data just like switching between email providers.

maybe the powers that be are made uneasy by the fact that these genetic tests will show that large numbers of white americans have some small amount of nonwhite ancestry.

One way the corporations get richer in america is that they are able to import large amounts of cheap nonwhite foreign labor. That importation is in large part due to the "white race guilt" propaganda drilled into white people from youth. If large numbers of whites find out they have traces of nonwhite ancestry, that "white race guilt" propaganda will have less effect, which means eventually the people will be less willing to give their consent to mass immigration. Less profit for the rich, then.

Suggest you take the National Geo Genographic test.

If you live in America you are an African American.

I suspect that what many people fear about DNA testing are the same things people fear when they start digging into their ancestry.

Even still, people should have that choice.

I'm curious...what's your position on ObamaCare? Safe info or not safe info?

I have had my genome sequenced (as part of the BGI project) but now what do I do with it?

Is there a service that analyzes/interprets genomes that have already been sequenced?

Is there some open archive where I can send it so it is available for people interested in research requiring large samples?

(I expect my genome is totally boring, but if it might be useful to somebody I'd b happy to add it to some big data bank.)

Take a gander at wmbriggs.com for more information on the stupidity of this process. The epigenome data causes the DNA data to be really damn murky.

There are instances where doing the gene thing makes sense. Most people though are not very good at dealing with the negative data.

I think your reasoning is kind of retarded. I've used the service, and found everything about it fascinating.

1) You forgot to use it. Just admit it.

2) It's pretty clear to take the information with a grain of salt. It didn't flag up anything significantly negative for me that I didn't already know (It pinpointed increased risk for two ailments that I already had family history of.) And if I had false-negatives, then not knowing I am in the same state I was prior to the test in not knowing I am susceptible. I found a lot of the other markers it came up with to be fascinating to read through, and it was surprisingly accurate (blood type, hair/eye color, size, balding, etc.)

3) By your reasoning, people will drive safer when they aren't wearing a seat belt.

4) It's just another notch in the warehouse of information people already have on you. Do you use this reasoning that someday your medical records may be leaked? Do you use this reasoning to avoid using credit cards because your purchase history may be leaked? Honestly, the leaking of my genome is probably the LEAST of the information I'd want leaked - my browsing history and purchasing history being far more embarassing and potentially career killing (at least for me) than someone knowing I have a family history of ass-cancer (which is already public knowledge if you follow my ancestry).

Also, I ended up communicating with a direct relative (first cousin) whom I never spoken with before. Which I thought was pretty cool in itself.

Last year I bought 23andme as a holiday gift for others who told me they were interested but had similar concerns to #2 and #3 and didn't want to get tested myself. Learning that the service might be neutered or dissolved got me off the fence - I just spit into the tube this morning.

What changed? Though technology will continue improving, the value of the 1m datapoints in the current service will also improve. As the functional consequence of individual SNPs (or combinations of SNPs) come into clearer focus, the data will become more informative. Having 1m SNPs genotyped for $99 is a fantastic value.

There is also an interesting question of potential costs of choosing *not* to get 23andme testing. If 23andme goes away and is replaced with testing that has to be ordered by a physician, what data goes into your personal health record? What about potentially negative incidental findings? It might be good for your doctor to know about them, but what about insurance companies? For me, I'd rather be in control of that data rather than having it part of my medical record.

Personal genomics isn't going away. Figuring out how to cope with the information is not going to be easy - particularly when that information reveals something unexpected or negative. 23andme starts, but by no means ends, this conversation.

I got the kit after reading Alex's post. I'm a little concerned about the privacy risks, but I'm also curious about the results so I went for it.

Regarding point 2), Tyler's point is probably valid for most people but there are exceptions. For example, women on my father's side of the family run a high risk getting ovarian cancer at an early age. My paternal grandmother, my paternal aunt, and my sister all suffered from it. Researchers have identified a specific genetic mutation which they believe is responsible. My female cousins were able to get tested for the gene and those that had the mutation opted for the most part to get their ovaries removed after they had finished having kids. Those that did not have the mutation were able to forgo this surgery with some peace of mind; in the absence of a genetic test they presumably would have been strongly tempted to have the surgery to avoid the serious risk entailed by this typically incurable disease.

Very interesting case. Thanks for sharing your story.

I would definitely be very interested in complete genome or even exome but the SNPs information that 23andme provides is simply too limited to pay more than $20 for some amusement. No one in my family has any simple enough hereditary diseases to worry about.

Comments for this post are closed