I would use that service. But, I would never, ever send them my DNA attached to my identity.
Sounds like I can't be the first to use Bitcoin with them either, come to think of it..
Your DNA IS your identity, and sending it to them is sending them all the information ever needed to identify you. For an encore you are also helping to identify your parents and your children and other relatives to a lesser degree depending on how far they are removed from you in the family tree.
That there is a name and a social security number attached to your DNA right when you send it is a convenience, not having that information is not an obstacle if any of your family members are also 23andme customers.
The only way this would work is if everybody sent them anonymous samples and 23andme would destroy each and every sample after receiving and processing it and destroy any and all records created as a result of processing that sample. No way they'll do that, and you could easily argue the only reason they exist is to build up that database.
I wrote about this a couple of years ago in some more detail:
http://jacquesmattheij.com/your-genetic-information-is-not-j...
Identical twins share DNA, but not identity.
But it has great potential to identify a person.
For example when you do not use an unique password (but perhaps falsify other personal information) for two separate user accounts.
Then your password can be used to relate two separate accounts together. This kind of interlinking is common practice in the industry for fraud detection and could be used for more sinister purposes.
But lets assume that you actually share your password with your mother and only flip some number of random bits in it (you can see an analogy with DNA).
Now assuming that your mother identity and password is known and you are only child in the family, your identity is also known given your password because there is fixed change distance between your and your mother password.
So I would say based on this that actually DNA is less unique than a fingerprint (assuming that there is no detectable correlation between parent and child fingerprint).
> Your DNA IS your identity, and sending it to them is sending them all the information ever needed to identify you.
Not really: in the case the cops already have the DNA but it's useless to them because they don't know how to find the person it belongs to. If the cops found an (erroneous) match to sequence done anonymously, they would learn little to nothing that they could act on.
Now the game might be up if they got a family-member match on a non-anonymous relative, but there's little you can do about that.
Of course you can. As parent above mentioned - you do not send your DNA. Plain and simple.
For some reason I had impression that it is about checker not having access to some ones DNA sample.
But someone with access to many sequence databases? Maybe they could identify some relatives, at least.
I'd like to impute more health and longevity goals to 23andme than just imagine it's a giant government DNA-gathering corporation. But, I bet you're right -- it's just not tenable to hide enough information about myself AND all my family members that I couldn't be ID'ed by a junior data scientist / low-end neural network.
There's also already some research with positive results about deanonymizing DNA:
http://www.sciencemag.org/content/339/6117/262.short
https://petsymposium.org/2015/papers/07_Humbert.pdf
The former is based on trying to use known data about family relationships to figure out who someone could be, while the latter is based on trying to figure out what someone would look like (or what other observable properties they could have) based on their DNA and then recognize them in a database of anonymous DNA samples.
While both of these studies have pretty significant constraints on the ability to do this kind of deanonymization in a general way for large databases, I think the power of DNA deanonymization will only grow as DNA databases and publicly available family tree records get larger, and as people learn to predict more and more phenotypic traits from a DNA sequence.