6 ms·

As someone who paid for 23andme+, it was pretty much a scam anyway, I got maybe 4-5 more reports over the year, all of which were for random popular ailments li

by belval 3y ago

As someone who paid for 23andme+, it was pretty much a scam anyway, I got maybe 4-5 more reports over the year, all of which were for random popular ailments like "Anxiety" where the link between your genes and the disease(?) are pretty dubious in the literature and the takeaway was "you are 5% more likely to experience anxiety".

astrange 3y ago
> where the link between your genes and the disease(?) are pretty dubious in the literature This is inescapable; genetics usually can't show causation, for instance because you can't do an experiment where you change someone's genes. Geneticists seem to deal with this by using statistics like GWAS that are obviously just correlation, adding a sentence that correlation doesn't show causation, and then just proceeding and pretending like it does.
- belval 3y ago
  Sure, perhaps that was a bit too succint. What I mean is that, as a customer of 23andme, I'd much rather know about some known genetic disease that I may or may not have than about wide-ranging things that affect everyone like stress and anxiety. It's just not as serious. 23andme went with the "only give very limited information" which I respect but is ultimately pretty much useless compared to just running your data against dbSNP and telling me that I have this or that marker for this or that disease backed by this or that paper.
- Dylan16807 3y ago
  Only showing correlation is one thing, but if the correlations themselves are barely noticable then that's a big problem for making a useful report. Though if you find whatever gene is most correlated with something, what are the options for it not to be causation? If the chance of causation is high enough, it makes sense to proceed as if the risk is real.
  - astrange 3y ago
    When I said correlation isn't causation I meant it. Neither high nor low correlation is evidence for causation. There's a more advanced form of being bad at this where you think you can show causation by controlling for everything in the environment. This is also wrong; it produces something called collider bias. > Though if you find whatever gene is most correlated with something, what are the options for it not to be causation? 1. It's a coincidence and it's never causal. Imagine an OSS project releases a bugfix and you diff the old and new versions. (This is basically GWAS.) The bugfix part of the diff caused it to be fixed. The updates to the copyright dates or changelogs didn't. 2. It's causal, but the causal chain involves a specific environmental factor, and we should change that instead. For instance, you can say every human has a genetic disease that prevents them from producing their own vitamin C, which most other mammals can do. But instead of calling scurvy a genetic disease we just eat fruits and vegetables.
    - Dylan16807 3y ago
      > 1. It's a coincidence and it's never causal. Half the point of analyzing statistics is to filter out coincidences, and that applies to correlations too. If something is a coincidence, it won't hold up as a proper correlation under reasonable amounts of analysis. So when the premise is we're starting with correlations, I think it's alright to assume they're mostly not coincidences. > The bugfix part of the diff caused it to be fixed. The updates to the copyright dates or changelogs didn't. In that case the bugfix is the "gene most correlated", isn't it? Give it a few generations to randomly spread, and the signal will be far stronger on the bugfix gene than on the copyright gene. (And if it hasn't been spreading for generations then you won't have enough samples to find either gene.) > 2. It's causal, but the causal chain involves a specific environmental factor, and we should change that instead. > For instance, you can say every human has a genetic disease that prevents them from producing their own vitamin C, which most other mammals can do. But instead of calling scurvy a genetic disease we just eat fruits and vegetables. If you're testing just humans, you'll get a 0% correlation because everyone has that gene. If you're testing across mammals, then "WARNING: Prone to scurvy". Which is a completely correct and causal result about a genetic problem, with easily accessible treatments. So, I don't understand your example at all.
- justrealist 3y ago
  Please have some awareness of how little thought you are putting into dismissing a field where thousands of very smart people are working on solving the problem you describe. In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. Have maybe a small ounce of humility in this respect.
  - astrange 3y ago
    > Please have some awareness of how little thought you are putting into dismissing a field where thousands of very smart people are working on solving the problem you describe. Thousands of very smart people do a lot of dumb things. People still work on string theory. The people working on proper causal inference are also smart, work hard, and have Nobel Prizes. https://www.nobelprize.org/prizes/economic-sciences/2021/popular-information/ https://www.nobelprize.org/prizes/economic-sciences/2021/pop... On the other hand, the people doing genetics who aren't careful about it produced the 23AndMe report which says I have a "17% chance of having green eyes". Against what counterfactual? > In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. "Correcting" is a wrong way to think about it. Generally speaking, overcorrection is worse than undercorrection because of collider bias. You need to choose a study design that's correct in the first place. You're describing a natural experiment, which is better than a GWAS of self-selected 23AndMe customers, but does have problems (silly one: selection bias because the sample only includes people with siblings) and more importantly is only guaranteed in the study environment (eg people who live in the UK in 2008) but gets reported without identifying what that environment is.
    - ImPostingOnHN 3y ago
      > > In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. > You need to choose a study design that's correct in the first place. That's precisely what a sibling study is: a study design that's correct in the first place, as far as the problem we're discussing goes. > You're describing a natural experiment That is correct: a sibling study, which corrects for the problem we're discussing, often contains aspects of a natural experiment, namely that the siblings tend to naturally be siblings, rather than raised together purely for the purpose of a study :)
KMag 3y ago
"10% more likely to experience anxiety, once having clicked on this useless link"

by belval 3y ago

astrange 3y ago

> where the link between your genes and the disease(?) are pretty dubious in the literature This is inescapable; genetics usually can't show causation, for instance because you can't do an experiment where you change someone's genes. Geneticists seem to deal with this by using statistics like GWAS that are obviously just correlation, adding a sentence that correlation doesn't show causation, and then just proceeding and pretending like it does.

belval 3y ago
Sure, perhaps that was a bit too succint. What I mean is that, as a customer of 23andme, I'd much rather know about some known genetic disease that I may or may not have than about wide-ranging things that affect everyone like stress and anxiety. It's just not as serious. 23andme went with the "only give very limited information" which I respect but is ultimately pretty much useless compared to just running your data against dbSNP and telling me that I have this or that marker for this or that disease backed by this or that paper.
Dylan16807 3y ago
Only showing correlation is one thing, but if the correlations themselves are barely noticable then that's a big problem for making a useful report. Though if you find whatever gene is most correlated with something, what are the options for it not to be causation? If the chance of causation is high enough, it makes sense to proceed as if the risk is real.
- astrange 3y ago
  When I said correlation isn't causation I meant it. Neither high nor low correlation is evidence for causation. There's a more advanced form of being bad at this where you think you can show causation by controlling for everything in the environment. This is also wrong; it produces something called collider bias. > Though if you find whatever gene is most correlated with something, what are the options for it not to be causation? 1. It's a coincidence and it's never causal. Imagine an OSS project releases a bugfix and you diff the old and new versions. (This is basically GWAS.) The bugfix part of the diff caused it to be fixed. The updates to the copyright dates or changelogs didn't. 2. It's causal, but the causal chain involves a specific environmental factor, and we should change that instead. For instance, you can say every human has a genetic disease that prevents them from producing their own vitamin C, which most other mammals can do. But instead of calling scurvy a genetic disease we just eat fruits and vegetables.
  - Dylan16807 3y ago
    > 1. It's a coincidence and it's never causal. Half the point of analyzing statistics is to filter out coincidences, and that applies to correlations too. If something is a coincidence, it won't hold up as a proper correlation under reasonable amounts of analysis. So when the premise is we're starting with correlations, I think it's alright to assume they're mostly not coincidences. > The bugfix part of the diff caused it to be fixed. The updates to the copyright dates or changelogs didn't. In that case the bugfix is the "gene most correlated", isn't it? Give it a few generations to randomly spread, and the signal will be far stronger on the bugfix gene than on the copyright gene. (And if it hasn't been spreading for generations then you won't have enough samples to find either gene.) > 2. It's causal, but the causal chain involves a specific environmental factor, and we should change that instead. > For instance, you can say every human has a genetic disease that prevents them from producing their own vitamin C, which most other mammals can do. But instead of calling scurvy a genetic disease we just eat fruits and vegetables. If you're testing just humans, you'll get a 0% correlation because everyone has that gene. If you're testing across mammals, then "WARNING: Prone to scurvy". Which is a completely correct and causal result about a genetic problem, with easily accessible treatments. So, I don't understand your example at all.
justrealist 3y ago
Please have some awareness of how little thought you are putting into dismissing a field where thousands of very smart people are working on solving the problem you describe. In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. Have maybe a small ounce of humility in this respect.
- astrange 3y ago
  > Please have some awareness of how little thought you are putting into dismissing a field where thousands of very smart people are working on solving the problem you describe. Thousands of very smart people do a lot of dumb things. People still work on string theory. The people working on proper causal inference are also smart, work hard, and have Nobel Prizes. https://www.nobelprize.org/prizes/economic-sciences/2021/popular-information/ https://www.nobelprize.org/prizes/economic-sciences/2021/pop... On the other hand, the people doing genetics who aren't careful about it produced the 23AndMe report which says I have a "17% chance of having green eyes". Against what counterfactual? > In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. "Correcting" is a wrong way to think about it. Generally speaking, overcorrection is worse than undercorrection because of collider bias. You need to choose a study design that's correct in the first place. You're describing a natural experiment, which is better than a GWAS of self-selected 23AndMe customers, but does have problems (silly one: selection bias because the sample only includes people with siblings) and more importantly is only guaranteed in the study environment (eg people who live in the UK in 2008) but gets reported without identifying what that environment is.
  - ImPostingOnHN 3y ago
    > > In fact, you can correct for the exact problem you describe with sibling studies where people have the same environment but different genes. > You need to choose a study design that's correct in the first place. That's precisely what a sibling study is: a study design that's correct in the first place, as far as the problem we're discussing goes. > You're describing a natural experiment That is correct: a sibling study, which corrects for the problem we're discussing, often contains aspects of a natural experiment, namely that the siblings tend to naturally be siblings, rather than raised together purely for the purpose of a study :)

KMag 3y ago

"10% more likely to experience anxiety, once having clicked on this useless link"