7 ms·

I have pitched to 23andme a few times over the last few years that they should accept Bitcoin, and offer totally PII-free purchases. This is a great use of a cr

I have pitched to 23andme a few times over the last few years that they should accept Bitcoin, and offer totally PII-free purchases. This is a great use of a cryptocurrency to my mind; a legitimate good which one should in no way attach to one's identity. 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..

schoen 11y ago
If the DNA services become popular enough, your identity might be deduced from your family members' DNA. There have already been a number of real-world stories where people used DNA in unexpected (and legally questionable) ways to identify people or relationships, including using family members' DNA as a starting point. There's also already some research with positive results about deanonymizing DNA: http://www.sciencemag.org/content/339/6117/262.short http://www.sciencemag.org/content/339/6117/262.short https://petsymposium.org/2015/papers/07_Humbert.pdf 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.
jacquesm 11y ago
> But, I would never, ever send them my DNA attached to my identity. 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-just-yours-but-it-is-family-property http://jacquesmattheij.com/your-genetic-information-is-not-j...
- dragonwriter 11y ago
  > Your DNA IS your identity Identical twins share DNA, but not identity.
  - jacquesm 11y ago
    For every rule there are exceptions, the number of identical twins is so low that for most practical purposes it does not matter. Unless you're a magician (see 'the prestige').
    - funkaster 11y ago
      that doesn't make any sense. DNA is not one's identity as in the metaphysical sense of a whole "self". DNA is just like fingerprint but more complete and more (?) unique. By your argument, your picture + some unique number could be your identity -> your driver's license is not your identity. It's a way to identify your identity.
      - eveningcoffee 11y ago
        So is my password my identity? Well, of course not. 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).
- tokenizerrr 11y ago
  In that case you also send over your entire identity any time you touch something, or even when losing some skin or hair.
jellicle 11y ago
Right. You'll use bitcoin! They won't know anything about you, except your gender, eye color, hair color, and that you're definitely related to Max Schubert, Adelaide Schubert, Jane Schubert, and Bob Schubert, all of whom reside in Little Rock, Arkansas and who did the "normal" DNA kit a few years ago, or a few years from now.
superuser2 11y ago
At the very least, package metadata would reveal the approximation location you sent the sample from.
slvrspoon 11y ago
you can use a service like Abine's Blur to signup, pay with a proxy credit card that never reveals your identity and pseudoynm and effectively be "anonymous". More here: https://www.abine.com/blog/2013/23andme-without-violating-your-genetic-privacy/ https://www.abine.com/blog/2013/23andme-without-violating-yo...

schoen 11y ago

If the DNA services become popular enough, your identity might be deduced from your family members' DNA. There have already been a number of real-world stories where people used DNA in unexpected (and legally questionable) ways to identify people or relationships, including using family members' DNA as a starting point. There's also already some research with positive results about deanonymizing DNA: http://www.sciencemag.org/content/339/6117/262.short http://www.sciencemag.org/content/339/6117/262.short https://petsymposium.org/2015/papers/07_Humbert.pdf 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.

jacquesm 11y ago

> But, I would never, ever send them my DNA attached to my identity. 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-just-yours-but-it-is-family-property http://jacquesmattheij.com/your-genetic-information-is-not-j...

dragonwriter 11y ago
> Your DNA IS your identity Identical twins share DNA, but not identity.
- jacquesm 11y ago
  For every rule there are exceptions, the number of identical twins is so low that for most practical purposes it does not matter. Unless you're a magician (see 'the prestige').
  - funkaster 11y ago
    that doesn't make any sense. DNA is not one's identity as in the metaphysical sense of a whole "self". DNA is just like fingerprint but more complete and more (?) unique. By your argument, your picture + some unique number could be your identity -> your driver's license is not your identity. It's a way to identify your identity.
    - eveningcoffee 11y ago
      So is my password my identity? Well, of course not. 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).
tokenizerrr 11y ago
In that case you also send over your entire identity any time you touch something, or even when losing some skin or hair.

jellicle 11y ago

Right. You'll use bitcoin! They won't know anything about you, except your gender, eye color, hair color, and that you're definitely related to Max Schubert, Adelaide Schubert, Jane Schubert, and Bob Schubert, all of whom reside in Little Rock, Arkansas and who did the "normal" DNA kit a few years ago, or a few years from now.

superuser2 11y ago

At the very least, package metadata would reveal the approximation location you sent the sample from.

slvrspoon 11y ago

you can use a service like Abine's Blur to signup, pay with a proxy credit card that never reveals your identity and pseudoynm and effectively be "anonymous". More here: https://www.abine.com/blog/2013/23andme-without-violating-your-genetic-privacy/ https://www.abine.com/blog/2013/23andme-without-violating-yo...