23andMe Genetic Testing Kit Free with a Year Long Membership of $108 | Hacker News Reader

23andMe Genetic Testing Kit Free with a Year Long Membership of $108

yourgeneticgenealogist.com

71 pointsrancar251 comments

My advice to anyone thinking of doing genetic testing (be it 23andMe or another route) is to consider seriously doing it at as a "John Doe" (ie not using your real name and details).

Knowing you have a high susceptibility to a significant disease could have all sorts of implications for insurance - medical, life and even car.

In general insurance companies require you to disclose any and all information that you have that would be pertinent to them assessing risk. Clearly for medical and life insurance you knowing there is a high chance you will get Parkinsons (for example) is information your insurer would like to know - and presumably will have a negative impact on your coverage costs/availability.

Here in the US there are currently laws - such as Genetic Information Nondiscrimination Act (GINA) - to prevent insurance companies demanding this information.

However laws can be repealed. The insurance industry is the 'leader' in government lobbying. 23andMe is still subject to US federal subpoenas laws. 23andMe could be aquired by an insurance firm.

Also consider laws differ in other countries, where insurance companies might be able to legally demand results. In Canada insurance companies can not only request it, they can demand you get this kind of testing before you can obtain coverage.

Consider further that databases can be hacked/stolen (it seems to be happening to every email marketing company right now).

Think carefully whether you want your personal legal name and contact details all over the results of a test like this.

if you expect to be truly anonymous with 23andme, you're going to need to anonymize the address the kit is sent to, the payment information, and the connections to other people you have as well.

"To date, there is no way to order the service completely anonymously. That said, if you have a trusted friend or relative who would be willing to order a kit for you, his or her address and credit card would be tied to the account, rather than your own, and there would be no way to connect it to you."

even if you've ordered under your own name, doesn't this suggest you could just maintain plausible deniability? surely a DNA test can connect it to you- but that would nullify the "john doe" tactic as well, no?

23andMe make clear that they do not assume that the genetic material belongs to you just because you are the account holder -- you can submit several samples under different names for family members, for instance.

It seems the simplest step to prevent 23AndMe being forced to give up your genetic data at some future dystopic point has already been taken: they already have no way of knowing if it's actually your data. You can take anyone's DNA and call it yours, or give anyone else's name to your own DNA.

The following is my boiler plate for anything that requires anonymous subscription - 23andme, domain registration, hosting, etc:

1) Mailbox in SF's China Town (a number don't ask for ID even though they're supposed to legally request it). Many also don't speak much English so if anyone comes to investigate good luck to them.

2) Visa/Amex gift card purchased at Walgreens/Safeway in cash - ideally in a different city you happen to be visiting with no connection in. Las Vegas and New York are great choices.

3) Email address created and accessed via TOR within the incognito mode of the browser. Buy the 23andMe subscription/domain/hosting in the same way but making sure you are using a different TOR session.

While I approve of you guarding your privacy, you may want to be careful about doing this with domain names - many registrars require that the WHOIS information is up to date, and can theoretically take them back if it isn't. I'd imagine the same holds for domainsbyproxy and the like.

So what? All you need is the initial email to work, nobody is going to notice fake whois that quickly. No person is going to show up at the address you list to verify that it's actually your address.

I've used fake whois records for a while and have never had any problems. I did make my main domain "real" a few years ago to avoid any problems, however.

Let's say someone wanted to buy as a John Doe. How would they pay without the risk of revealing their identity?

see http://news.ycombinator.com/item?id=2431482

I bought my entire family the $99 special on last year's 23andme. The kits are pretty darn anonymous - and they all signed up under initials. Unless you give them your legal name, their is no way to identify who the test was for - address doesn't tell you that.

If 23andMe is acquired by any company (insurance or otherwise) that wants to change the Terms Of Service and / or Privacy Policy, they will have to notify customers, and those customers who don't agree with the changes would have the right (under the current policy) to download their raw data and delete their 23andMe accounts.

Also, the chances of 23andMe getting hacked are much lower than the chances of your friendly neighborhood email marketing company. I know since I built the site.

Is that your aspiration or something you know for a fact?

I couldn't see anything that indicates this in their Privacy Policy. In fact https://www.23andme.com/about/privacy/#Full says:

Account Closure and Correction of Personal Information

As stated in the applicable Consent Document, however, Genetic Information and/or Self-Reported Information that you have previously provided and for which you have given consent to use in 23andWe Research will not be removed from ongoing or completed studies that use the information. Our contracted genotyping laboratory may also retain your Genetic Information as required by local law and we may retain backup copies for a limited period of time pursuant to our data protection policies. In addition, we retain limited Registration Information related to your order history (e.g., name, contact, and transaction data) for accounting and compliance purposes.

Doesn't look like they're interested in totally zero'ing you out of all of their records even if you explicitly request it (acquisition or not).

From the page you cited (https://www.23andme.com/about/privacy/#Full):

Business Transitions

In the event that 23andMe goes through a business transition such as a merger, acquisition by another company, or sale of all or a portion of its assets, your Personal Information will likely be among the assets transferred. In such a case, your information would remain subject to the promises made in any pre-existing Privacy Statement.

From their current privacy policy (https://www.23andme.com/about/privacy/?version=1.1):

Business Transitions

In the event that 23andMe goes through a business transition such as a merger, acquisition by another company, or sale of all or a portion of its assets, your personal information and non-personal information will likely be among the assets transferred. You will be notified in advance via email and prominent notice on our website of any such change in ownership or control of your personal information. We will require an acquiring company or merger agreement to uphold the material terms of this privacy statement, including honoring requests for account deletion.

I think khomenko is correct, with the exception you pointed out (ongoing studies).

It is something I know for a fact since I worked at 23andMe for 4+ years. Looks like the first responder already found the relevant section and posted it. It's not an ironclad assurance, and there are a few caveats, but these provisions, together with any scrutiny an attempted acquisition would bring (the database is big and well-known enough where it's virtually guaranteed), can provide peace of mind for some. There are certainly risks and I don't advocate that everyone gets tested without considering them, but "insurance company buys 23andMe, gets all the data, and starts discriminating based on it", is a very low-probability event based on what I know.

It says that I would have to pay 9 dollars a month. Do you know if I can tell them to charge the whole subscription once instead of 9 dollars every month?

You can "buy it as a gift" and pre-pay the 1-year subscription.

Damn, I already paid, is there any way I could do it now or it's too late?

Contact help@23andme.com

if the issue is knowing whether you knew about a disease, i don't think there's any way to anonymize oneself.

the insurance company could simply acquire a dna sample from you and check whether there is a match in the database?

In theory it could have been someone else who sent in your sample.

On the other hand, it presents an opportunity in countries with a moratorium on using genetic testing information for insurance purposes; it creates an information asymmetry in your favour. For example, until 2017 in the UK you can take out as much life insurance as you please with no test-result penalty (with the sole exception of Huntington’s Disease results). http://www.abi.org.uk/Media/Releases/2011/04/Insurance_Genet... Perhaps the insurance companies figure that knowing the results might improve the actual outcome in many cases, because the individual alters their behaviour to mitigate them. That, and the predictive power of most tests is fairly weak at present.

https://github.com/msporny/dna

a hackernews reader: https://github.com/orta/dna

What's the chance of this kind of thing occurring and simultaneously NOT requiring a new test anyway?

I know that it's entirely irrational and counter-productive, but I can not bring myself to do this and find out things I'd rather not spend the rest of my (hopefully not too short) life obsessing over. I'm already busy enough focusing on the obvious things, like being overweight. If I discovered that I'm certain to end up with Parkinson's or Alzheimer's or something, I might have a pretty difficult time coping.

It makes no sense. It's entirely irrational. I love information. Knowledge. Informed decision making. And yet . . .

Believe me, it's unlikely you'll get anything that dramatic.

This.

I don't even like going to Web MD, in fear of exaggerating a condition or, even worse, synthesizing symptoms. I can't imagine knowing several years ahead of time that I have a high chance of developing this or that unpreventable disease. I'd probably end up seeing "signs" long before they ever presented themselves or perhaps even worsen my condition through the anxiety.

It's not irrational at all. If there's not yet any pragmatic benefit to knowing that you have a significant risk of getting a serious disease in the future, it's probably better not to know.

You can select which tests to make available to yourself, though, and avoid finding out about things like Parkinson's.

Even though there's no current treatment for early-onset Alzheimer's, I'd certainly want to know if I was at risk. I'd live my life much differently if I thought my brain would rot at age 50.

Not knowing if you'll get an untreatable disease doesn't stop you from getting an untreatable disease! It just means it'll sneak up on you and force you to make tough decisions with less preparation.

"I'd live my life much differently if I thought my brain would rot at age 50."

What would you be doing so differently? I'm really just curious.

There's more to life than creating a successful start-up.

It would change my career and retirement planning quite a bit. I'd retire early. I would forget having kids, since I wouldn't be around to support them after their mid-teens.

Think about it from the flip side: If you knew you were going to live to age 200 in good health, wouldn't you retire later? Maybe you'd spend more time gallivanting around the world like a recent college grad. Maybe you would wait longer to settle down and have children. Any deviation from the current typical life-span should definitely change your behavior.

"I'd live my life much differently if I thought my brain would rot at age 50."

Not to get all Carpe Diem on you, but I'd suggest if this is true, you should probably re-evaluate your life. Whether your brain starts to rot at age 50, or you get hit by a bus in two weeks, you don't get to control how much time you have. You can only control what you do with that time.

If you live every day like it's your last then that precludes sensible things like retirement planning. On the other hand, if you know you're not going to be mentally capable of enjoying your retirement anyway, then now is the time to blow that cash.

I'm not aware of very much in the genetic material 23andme processes that is indicative of a 100% chance of having some future disease. Here's a screenshot of my 23andme dashboard with some of the information removed: http://i.imgur.com/9ViAf.jpg . As you can see, everything here is based on how much more likely you may be to develop a disease than the average person. Within the page details the specific studies that have given them the information being used in the calculation are cited with links and additional technical information. You can also see that the highest risk factor they've calculated for me is 11.3% or 1.61x the average (in this case, it's Age-related Macular Degeneration).

Obviously, it's possible that a person could be given a more significant percentage of a serious condition which raises additional questions. If you are both A) objective enough to firmly grasp that a chance is simply a chance, not a sure outcome and B) not someone who worries a great deal about possible future situations, you might benefit from using 23andme. Otherwise, it may be wise not to purchase the program.

For me, I've had an interesting effect of using some of the increased negative risks to help me stop bad habits while at the same time, not using any of the decreased risks as a license to go ahead and live more recklessly in those areas. Hopefully this give you some additional perspective on whether something like this would be right for you.

That's actually a great example of how the information could be helpful.

If I knew I had a high-risk of Macular Degeneration, I might take some time now to learn Emacspeak or any of the other screen-reader technologies so that I could still do software development in the event that my vision left me.

be careful, very easy to decode your health information from this screenshot.

( the average risk - is mostly unique value )

( for example: Restless Legs Syndrome [ avg risk: 2.0% ])

:-)

Yes, I realized this when making the image. It's certainly something people should consider before posting their results online even in an edited format. For me, blocking the names out was a barrier to discovery for most people (not worth the effort). It's not something I specifically want to share with the world, but there's nothing in that list I would be upset to hear people know either. Certainly a good comment for people to think about.

I feel the same way, but there are benefits too. If you know you are high risk for Parkinson's you can get screened earlier and potentially start treatment earlier.

And you can do things known to lower your risk, like start drinking caffeine every day.

You've acknowledged it's irrational, so making rational points in rebuttal is clearly a pointless exercise. Nevertheless...

If you are already worrying about "obvious" things, wouldn't it be better to know whether or not you have anything to worry about? For instance, if it turns out you have a below-average chance of heart disease, it might make you worry less about your weight. On the other hand, if you discovered that you were (like me) at elevated risk of developing diabetes, wouldn't that provide the extra motivation you need to limit your suger intake and further control your weight?

I can't see any situation where having less data would make things demonstrably better -- just more uncertain.

You are not alone in your fear.

This question of knowing your fate was well presented in Heinlen's short story Life-line: http://en.wikipedia.org/wiki/Life-Line

While I like the idea and being trailblazers must be done by someone you have to keep in mind that 23andMe only do genotyping not full sequencing.

In my mind the difference between genotyping and full sequencing is between building a partial array from bunch of if statements and actually copying the array.

I paid for 23andme last year during a sale. I have a few markers that are in the 'you'll probably get that' category and as of yet, I haven't stressed out about any of them.

However, if I go to the doctor and he says I'm at risk for environment factors for that, I'll know there's some serious shit to think about.

And it could save my life!

If you understand probability, then the data on 23andme will not destroy your life. It'll merely give you an additional warning sign of things to really pay attention to.

And if you choose, you can ignore it altogether.

For anyone who missed the details on when this is available (as I did on the first read) and who read it thinking of taking advantage of this offer: the deal is only available from midnight to midnight (24 hours) Monday, using PST as the timezone. So starting in just under 4 hours.

$63 to ship a single kit to The Netherlands? What is this, eBay?

Agreed - but it's medical mail both ways with import duty paid on way in... I bit the bullet and paid - not bad, about £100 all in for the year sub and test. Hopefully the data will be of interest.

One of the more interesting parts of 23andme, at least for me, was finding second and third cousins (likely) that I didn't know. Most of the actual medical tests are fairly non-predictive (1.5x increase in one study for something which is already rare?), but the ancestry part is something where having a lot of participants in a selectively-open database can be interesting.

I do really hope Halcyon (or another company) gets full sequencing costs down as low as they say they can. Then, sequencing becomes just a routine checkbox on blood tests.

Note that the extent of their genetic analysis varies by ethnicity, with those of East Asian and African descent getting less than half as much analysis: https://www.23andme.com/health/ethnicity/

That's certainly true right now, but the information in the health reports gets updated all the time based on new studies, so it will continue to get better.

"Free when you pay"

I agree. It was a word choice between free and sale. I assumed free was a more effective title than sale for HN.

Note that when buying add a one-year subscription which will bring the total price to $108.

Just told someone to order, i think the service is worth every penny, i hope doctors find this research valid enough to base some of their work on it(i'm public insured in germany).

My only fear at this point is them turning people over to evil scientists after finding out about their superpowers.

Drats...largely limited to NA and Europe.

Reply on news.ycombinator.com