5 ms·

23andMe offers a very low cost way to open the door to a whole new world of having people understand something about their genetic makeup and the possible impac

by crusso 13y ago

23andMe offers a very low cost way to open the door to a whole new world of having people understand something about their genetic makeup and the possible impact upon their lives. The work that they're doing to show how genetic features correlate with the observed tastes, behaviors, and propensities of users of the site seems remarkably useful. The FDA is doing exactly what it always does, stifling innovation while justifying its existence using the voluminous laws created by politicians justifying their existences. If the FDA were useful, it would be helping 23andMe rather than trying to take them down.

rayiner 13y ago
> The work that they're doing to show how genetic features correlate with the observed tastes, behaviors, and propensities of users of the site seems remarkably useful. Oh man are you going to be upset when we regulate all that out of existence. Trying to correlate genetics to tastes, behaviors, and propensities, while scientifically interesting, is a total political non-starter. > The FDA is doing exactly what it always does, stifling innovation while justifying its existence using the voluminous laws created by politicians justifying their existences. This action is the FDA acting within the core of its jurisdiction, and within the core of a law that was passed with tremendous public support. This is not to say that agencies don't do what you're describing, but in this case, it's pretty much the FDA's job to control this sort of "innovation."
- masklinn 13y ago
  > while scientifically interesting I'm not sure "interesting" is the word I'd have used. "Unlikely" is probably much, much closer to the truth. It's more or less astrology.
  - crusso 13y ago
    Correlating genetics to propensities is astrology? So, correlating the existence of certain genes to propensity of heart disease is mumbo jumbo? Not really sure why anyone would say such a thing. Maybe you didn't understand what the word "propensities" means?
    - anigbrowl 13y ago
      When you put it in the same sentence as 'tastes and behaviors' you're creating a very different context from things like heart disease. Imagine, for example, a genetic test that says that you, Crusso, are 29% more likely to commit a criminal act involving deception than the median person. Now imagine that your genetic testing service has a data breach and suddenly you have immense difficulty getting hired, getting credit, or getting a table in a restaurant. One has only to glance back to the last century to see overconfident claims along the same lines from eugenicists, for example. Or look at the dreadful state of science journalism and the broken logic that prevails in many internet comment/forum sections to see how easily a statistical correlation can be substituted for fact - to say nothing of poor statistical work in scientific research, where publication strongly favors positive results and replication is often under-prioritized. Now please return to your preventive custody zone, non-citizen.
      - crusso 13y ago
        statistical correlation can be substituted for fact Well let's stop this Science nonsense and get on with full blown Political Correctness as our guide to the future we're heading toward.
        masklinn 13y ago
        That's got nothing to do with PC or any such nonsense, it's got to do with you taking as your starting point that it's possible to reduce highly complex, multivariate and environmentally influenced phenotypes down to useful genotypes. Your comments are essentially daily mail fare.
        Dylan16807 13y ago
        anigbrowl's comment argued that correlations were bad, and did not mention whether useful ones were possible. Blame them for bringing up political correctness, not crusso.
        13y ago
- crusso 13y ago
  Oh man are you going to be upset when we regulate all that out of existence I guess you're being sarcastic, but it's not that I fear progress will stop. What I see instead is a great deal of needless friction that prevents efficiency. When you consider how innovations and businesses build on one another over the years - any friction in the process compounds the amount of time for "the future" to be here. I doubt anyone here on HN doesn't realize that in the future our genetic code will be understood and manipulated like we do with computers and software today. Will this golden age of genetic self determination be 50 years from now or 500 years from now? Roadblocks we put in front of progress now will mean many decades of delay for the advances down the road that are dependent upon today's advances and the intervening thousands of others. law that was passed with tremendous public support Doesn't at all make it right, good, or mean I won't bitch about it.
  - rayiner 13y ago
    > I guess you're being sarcastic, but it's not that I fear progress will stop. I'm not being sarcastic. Any technology that tries to correlate genetic markers with behavioral characteristics will be regulated out of existence. I'm not taking a position one way or the other as to whether this is a good idea, but I'd bet money on it being the outcome. > Doesn't at all make it right, good, or mean I won't bitch about it. Sure. But what it does mean is that it's an intended consequence, rather than an unintended consequence as your characterization suggested.
    - crusso 13y ago
      Any technology that tries to correlate genetic markers with behavioral characteristics will be regulated out of existence I'm sure that there are those who would like to make all such pursuits illegal. Such a slippery slope, though. Where will that line be drawn? Which 23AM style correlation question should be illegal for inclusion in their correlation determination system? 1. Do brussel sprouts taste bitter and unpleasant? 2. Are you more active in the mornings? 3. Do you prefer blondes to brunettes? 4. Do you carry debt on your credit cards month-to-month or pay them off? 5. Did you graduate college? 6. Do your joints ache after rigorous exercise? 7. Have you ever been arrested for a felony? 8. Are people with a great deal of money more lucky than clever? 9. Does hard work tend to lead to financial success?
pilom 13y ago
No the FDA is saying that they need to see proof that the test has a high level of accuracy and isn't just snake oil. 23andMe tried saying "you shouldn't actually use our test for diagnosing disease" to which the FDA said "then stop marketing it that way." 23andMe both wants to _say_ they have a totally accurate test, but doesn't want to pay to do the testing to prove it is totally accurate. The FDA is saying you can't have it both ways.
- dnautics 13y ago
  have you actually used 23andMe? All over the place, it says, the information is based on academic studies, these figures are only n star (on a rating from 1-5) reliable, it only is conclusive for ethnicities X, etc, etc, etc. Now, it is not clear to me that people read all that (in the aldehyde dehydrogenase test, a lot of caucasians don't realize that it only applies to asians, even though it says that in bold, font right underneath the title). There is no such thing as a totally accurate test and I don't think 23andMe says that anywhere.
  - summerdown2 13y ago
    I'm not sure total accuracy is the point. Rather, I think it's about the claimed value being correct. If 23andMe say a test has a 20% accuracy, they should be able to prove how they came by that number. There's a big difference between telling someone they have a 5% chance of something and telling them they have a 25% chance of something - even though both are uncertain.
  - pilom 13y ago
    The issue isn't them saying "you have a higher likelyhood of disease." The problem the FDA has is 23andMe saying "you have the gene variation XYZ." If you take your 23andMe results to a doctor, the doctor should at best order a REAL FDA approved diagnostic test to see if you really do have variation XYZ because 23andMe hasn't proven that their test is accurate.
HarryHirsch 13y ago
If the FDA were useful, it would be helping 23andMe rather than trying to take them down. It's not their mandate to help 23andme, their mandate is to protect the general public. Their enforcement letter clearly spells it out: the FDA had had concerns about the validity of some of their tests, and they gave 23andme opportunity to address there. Since they failed, the FDA is taking action.
- nickff 13y ago
  The FDA protects the public, but only from new things, it does not protect new from a disease, but it does protect you from a potentially unproductive test or treatment. One could say the FDA exists to ensure that all statistical errors are of type II. http://en.wikipedia.org/wiki/Type_I_and_type_II_errors http://en.wikipedia.org/wiki/Type_I_and_type_II_errors
DanBC 13y ago
> If the FDA were useful, it would be helping 23andMe rather than trying to take them down. FDA has tried to help 23andme. 23andme declined the help.
metaphorm 13y ago
let me guess, you're an investor in 23andMe?
icegreentea 13y ago
How the hell should the FDA be trying to help 23andMe? Did you even read the letter? The FDA is making serious and legitimate claims. Amongst other things, the FDA is claiming that 23andMe has said that they were in the process of completing their validation studies. 23andMe has already submitted their 510(k) (which is the form/clearance required for -any- medical device), but failed to provide the required evidence to back up their claims in their submission. Note that the FDA is that claiming that 23andMe has not provided sufficient evidence (as in they don't like the study you did) rather the FDA is claiming that 23andMe has not submitted the required evidence at all. As in there should be this big ass binder full of test results comparing 23andMe's test results against the ground truth, but that binder doesn't exist. This is hardly stifling innovation. 23andMe has had 5 years since they started selling their product to produce an analytical or clinical study proving their product's effectiveness, and they failed to do so, despite clearly knowing that it was required, to the point of claiming that they were in fact working on it.
- dnautics 13y ago
  5 years for a clinical study? That's a laughably short amount of time. You don't just write a test suite and get your answer out. Moreover, this is kind of uncharted territory: If you've got a device that, say, estimates an 'elevated' risk of alzheimer's in 20 year olds, how long of a clinical trial period do you suspect would be reasonable to run the test?
  - icegreentea 13y ago
    5 years for an analytical/clinical study for a DNA test? Absolutely in the cards. 23andMe isn't discovering new mutations that it then claims are linked to some phenotypical difference. They are going through all the SNPs that have ever been studied by someone else, testing you for it, and then telling you that "You have this SNP which is linked to clopidogrel metabolism" for example. The study that the FDA would have required from 23andMe is to show that the genetic testing provides the correct results at the bare minimal. They would require 23andMe to quantify their actual error rate of their product in use by customers. Depending on 23andMe's claims in their 510(k) submission regarding their intended uses, then the burden of proof may go higher. The point is that if 23andMe can't actually prove that this test result predicts elevated risk of Alzheimer's (either by their own study, or through other relevant studies) then they sure as hell shouldn't say that their test predicts Alzheimer's risk. Case in point, clopidogrel response. The link between CYP2C19 and clopidogrel response is well studied. There are already other tests on the market based on this relationship. Given this, 23andMe would not have to prove by themselves that clopidogrel and CYP2C19 are linked, they would merely have to prove their their CYP2C19 test provided correct results.
dragonwriter 13y ago
> If the FDA were useful, it would be helping 23andMe rather than trying to take them down. If the FDA were trying to take 23andMe down rather than helping them, they wouldn't have been working with them for several years to get them into compliance and only issued a Warning Letter after 23andMe walked away from the process.
- HeyLaughingBoy 13y ago
  Exactly. The FDA takes this shit seriously and will order you to take products off the market or shut you down completely if you fail to comply with their orders. Abbot famously ran afoul of this 10 years ago and it looks like they're having trouble again. http://www.mddionline.com/node/5701 http://www.mddionline.com/node/5701

by crusso 13y ago

rayiner 13y ago

> The work that they're doing to show how genetic features correlate with the observed tastes, behaviors, and propensities of users of the site seems remarkably useful. Oh man are you going to be upset when we regulate all that out of existence. Trying to correlate genetics to tastes, behaviors, and propensities, while scientifically interesting, is a total political non-starter. > The FDA is doing exactly what it always does, stifling innovation while justifying its existence using the voluminous laws created by politicians justifying their existences. This action is the FDA acting within the core of its jurisdiction, and within the core of a law that was passed with tremendous public support. This is not to say that agencies don't do what you're describing, but in this case, it's pretty much the FDA's job to control this sort of "innovation."

masklinn 13y ago
> while scientifically interesting I'm not sure "interesting" is the word I'd have used. "Unlikely" is probably much, much closer to the truth. It's more or less astrology.
- crusso 13y ago
  Correlating genetics to propensities is astrology? So, correlating the existence of certain genes to propensity of heart disease is mumbo jumbo? Not really sure why anyone would say such a thing. Maybe you didn't understand what the word "propensities" means?
  - anigbrowl 13y ago
    When you put it in the same sentence as 'tastes and behaviors' you're creating a very different context from things like heart disease. Imagine, for example, a genetic test that says that you, Crusso, are 29% more likely to commit a criminal act involving deception than the median person. Now imagine that your genetic testing service has a data breach and suddenly you have immense difficulty getting hired, getting credit, or getting a table in a restaurant. One has only to glance back to the last century to see overconfident claims along the same lines from eugenicists, for example. Or look at the dreadful state of science journalism and the broken logic that prevails in many internet comment/forum sections to see how easily a statistical correlation can be substituted for fact - to say nothing of poor statistical work in scientific research, where publication strongly favors positive results and replication is often under-prioritized. Now please return to your preventive custody zone, non-citizen.
    - crusso 13y ago
      statistical correlation can be substituted for fact Well let's stop this Science nonsense and get on with full blown Political Correctness as our guide to the future we're heading toward.
      - masklinn 13y ago
        That's got nothing to do with PC or any such nonsense, it's got to do with you taking as your starting point that it's possible to reduce highly complex, multivariate and environmentally influenced phenotypes down to useful genotypes. Your comments are essentially daily mail fare.
        Dylan16807 13y ago
        anigbrowl's comment argued that correlations were bad, and did not mention whether useful ones were possible. Blame them for bringing up political correctness, not crusso.
        13y ago
crusso 13y ago
Oh man are you going to be upset when we regulate all that out of existence I guess you're being sarcastic, but it's not that I fear progress will stop. What I see instead is a great deal of needless friction that prevents efficiency. When you consider how innovations and businesses build on one another over the years - any friction in the process compounds the amount of time for "the future" to be here. I doubt anyone here on HN doesn't realize that in the future our genetic code will be understood and manipulated like we do with computers and software today. Will this golden age of genetic self determination be 50 years from now or 500 years from now? Roadblocks we put in front of progress now will mean many decades of delay for the advances down the road that are dependent upon today's advances and the intervening thousands of others. law that was passed with tremendous public support Doesn't at all make it right, good, or mean I won't bitch about it.
- rayiner 13y ago
  > I guess you're being sarcastic, but it's not that I fear progress will stop. I'm not being sarcastic. Any technology that tries to correlate genetic markers with behavioral characteristics will be regulated out of existence. I'm not taking a position one way or the other as to whether this is a good idea, but I'd bet money on it being the outcome. > Doesn't at all make it right, good, or mean I won't bitch about it. Sure. But what it does mean is that it's an intended consequence, rather than an unintended consequence as your characterization suggested.
  - crusso 13y ago
    Any technology that tries to correlate genetic markers with behavioral characteristics will be regulated out of existence I'm sure that there are those who would like to make all such pursuits illegal. Such a slippery slope, though. Where will that line be drawn? Which 23AM style correlation question should be illegal for inclusion in their correlation determination system? 1. Do brussel sprouts taste bitter and unpleasant? 2. Are you more active in the mornings? 3. Do you prefer blondes to brunettes? 4. Do you carry debt on your credit cards month-to-month or pay them off? 5. Did you graduate college? 6. Do your joints ache after rigorous exercise? 7. Have you ever been arrested for a felony? 8. Are people with a great deal of money more lucky than clever? 9. Does hard work tend to lead to financial success?

pilom 13y ago

No the FDA is saying that they need to see proof that the test has a high level of accuracy and isn't just snake oil. 23andMe tried saying "you shouldn't actually use our test for diagnosing disease" to which the FDA said "then stop marketing it that way." 23andMe both wants to _say_ they have a totally accurate test, but doesn't want to pay to do the testing to prove it is totally accurate. The FDA is saying you can't have it both ways.

dnautics 13y ago
have you actually used 23andMe? All over the place, it says, the information is based on academic studies, these figures are only n star (on a rating from 1-5) reliable, it only is conclusive for ethnicities X, etc, etc, etc. Now, it is not clear to me that people read all that (in the aldehyde dehydrogenase test, a lot of caucasians don't realize that it only applies to asians, even though it says that in bold, font right underneath the title). There is no such thing as a totally accurate test and I don't think 23andMe says that anywhere.
- summerdown2 13y ago
  I'm not sure total accuracy is the point. Rather, I think it's about the claimed value being correct. If 23andMe say a test has a 20% accuracy, they should be able to prove how they came by that number. There's a big difference between telling someone they have a 5% chance of something and telling them they have a 25% chance of something - even though both are uncertain.
- pilom 13y ago
  The issue isn't them saying "you have a higher likelyhood of disease." The problem the FDA has is 23andMe saying "you have the gene variation XYZ." If you take your 23andMe results to a doctor, the doctor should at best order a REAL FDA approved diagnostic test to see if you really do have variation XYZ because 23andMe hasn't proven that their test is accurate.

HarryHirsch 13y ago

If the FDA were useful, it would be helping 23andMe rather than trying to take them down. It's not their mandate to help 23andme, their mandate is to protect the general public. Their enforcement letter clearly spells it out: the FDA had had concerns about the validity of some of their tests, and they gave 23andme opportunity to address there. Since they failed, the FDA is taking action.

nickff 13y ago
The FDA protects the public, but only from new things, it does not protect new from a disease, but it does protect you from a potentially unproductive test or treatment. One could say the FDA exists to ensure that all statistical errors are of type II. http://en.wikipedia.org/wiki/Type_I_and_type_II_errors http://en.wikipedia.org/wiki/Type_I_and_type_II_errors

DanBC 13y ago

> If the FDA were useful, it would be helping 23andMe rather than trying to take them down. FDA has tried to help 23andme. 23andme declined the help.

metaphorm 13y ago

let me guess, you're an investor in 23andMe?

icegreentea 13y ago

How the hell should the FDA be trying to help 23andMe? Did you even read the letter? The FDA is making serious and legitimate claims. Amongst other things, the FDA is claiming that 23andMe has said that they were in the process of completing their validation studies. 23andMe has already submitted their 510(k) (which is the form/clearance required for -any- medical device), but failed to provide the required evidence to back up their claims in their submission. Note that the FDA is that claiming that 23andMe has not provided sufficient evidence (as in they don't like the study you did) rather the FDA is claiming that 23andMe has not submitted the required evidence at all. As in there should be this big ass binder full of test results comparing 23andMe's test results against the ground truth, but that binder doesn't exist. This is hardly stifling innovation. 23andMe has had 5 years since they started selling their product to produce an analytical or clinical study proving their product's effectiveness, and they failed to do so, despite clearly knowing that it was required, to the point of claiming that they were in fact working on it.

dnautics 13y ago
5 years for a clinical study? That's a laughably short amount of time. You don't just write a test suite and get your answer out. Moreover, this is kind of uncharted territory: If you've got a device that, say, estimates an 'elevated' risk of alzheimer's in 20 year olds, how long of a clinical trial period do you suspect would be reasonable to run the test?
- icegreentea 13y ago
  5 years for an analytical/clinical study for a DNA test? Absolutely in the cards. 23andMe isn't discovering new mutations that it then claims are linked to some phenotypical difference. They are going through all the SNPs that have ever been studied by someone else, testing you for it, and then telling you that "You have this SNP which is linked to clopidogrel metabolism" for example. The study that the FDA would have required from 23andMe is to show that the genetic testing provides the correct results at the bare minimal. They would require 23andMe to quantify their actual error rate of their product in use by customers. Depending on 23andMe's claims in their 510(k) submission regarding their intended uses, then the burden of proof may go higher. The point is that if 23andMe can't actually prove that this test result predicts elevated risk of Alzheimer's (either by their own study, or through other relevant studies) then they sure as hell shouldn't say that their test predicts Alzheimer's risk. Case in point, clopidogrel response. The link between CYP2C19 and clopidogrel response is well studied. There are already other tests on the market based on this relationship. Given this, 23andMe would not have to prove by themselves that clopidogrel and CYP2C19 are linked, they would merely have to prove their their CYP2C19 test provided correct results.

dragonwriter 13y ago

> If the FDA were useful, it would be helping 23andMe rather than trying to take them down. If the FDA were trying to take 23andMe down rather than helping them, they wouldn't have been working with them for several years to get them into compliance and only issued a Warning Letter after 23andMe walked away from the process.

HeyLaughingBoy 13y ago
Exactly. The FDA takes this shit seriously and will order you to take products off the market or shut you down completely if you fail to comply with their orders. Abbot famously ran afoul of this 10 years ago and it looks like they're having trouble again. http://www.mddionline.com/node/5701 http://www.mddionline.com/node/5701