3 ms·

Yes - (1) and (4) are extremely unlikely, though (2) and (3) are certainly plausible :) More specifically, this study may publish some marginal results about g

by bthomas 14y ago

Yes - (1) and (4) are extremely unlikely, though (2) and (3) are certainly plausible :) More specifically, this study may publish some marginal results about genetics of extreme IQ, but any single effect will likely be so small that nothing actionable will come from it. If (4) is the goal, I suspect they'll have much more success exploring drugs and other cognitive enhancements. What genetics can do is help inform how to discover those drugs - ie. what genes should you study. (Shameless plug: if you like genetics and writing python, drop me an email if you want to work on studies like this!)

Scaevolus 14y ago
Why do you think discovering genetic components to intelligence is extremely unlikely? I expect it to be easier to find genetic components for intelligence than for other polygenetic traits like height -- the brain has fewer components influencing development.
- bthomas 14y ago
  > the brain has fewer components influencing development What do you mean by that? I'd argue the opposite. In a sense, we already have discovered genes for high height with Marfan Syndrome. I think IQ is orders of magnitude more complex. Awesome conversation but probably hard in an online forum, so I'll just cite my favorite example: Dickens and Flynn [1] I absolutely think this study has a chance of revealing some signal, my skepticism is just that it will have any actionable conclusions. Here's a litmus test: what results would allow you to choose from a group of 10 men and 10 women, based only on genotype, which pair is most likely to have a high IQ child. And how would your power compare to just reading all their resumes? [1] http://en.wikipedia.org/wiki/Heritability_of_IQ#Dickens_and_Flynn_model http://en.wikipedia.org/wiki/Heritability_of_IQ#Dickens_and_...
  - Scaevolus 14y ago
    Barring specific diseases like Marfan Syndrome, we can't yet explain most of height -- "Height is a highly heritable trait, with estimates of heritability as high as 90%. Recent genome-wide association studies of height have discovered over 180 common variants associated with height. These variants have small effect sizes and collectively explain approximately 10% of the heritability." [1] Growth involves many systems with many types of cells interacting -- digestive, skeletal, endocrine, etc. There's a lot of complexity when you have so many systems involved. By comparison, there are relatively few distinct types of cells in the brain, so the effects of different alleles should be more apparent. In particular, genes controlling sensitivity to and production of neurotransmitters directly impact brain function. If they find many specific mutations that cause better performance on specific tasks (working memory, pattern matching, verbal ability, ...) it should be possible to predict high IQs. I don't expect 100% accurate predictions, but to not find _anything_ would be surprising. Maybe that's your point-- if they only get 10% heritability, it's not really enough for breeding decisions. [1] http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002439 http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fj...
  - JoshuaDavid 14y ago
    I recognize that your litmus test was rhetorical, but it's a question that does seem answerable, so I'll try to answer it. I would imagine there are a number of genes that are associated with lowered intelligence. For example, one might start by looking at genes such as the Dopamine D4 receptor gene, for which the 7-repeat allele seems to be a pretty good indicator of ADHD [1][2], which in turn is associated with significantly lower IQ. Of course, chains of correlation don't always work, and that is why it would be important to have tens or preferably hundreds to thousand of things like that to look for. While it would be nice and easy to look for an "intelligence gene", that may not be feasible. However, my guess is that looking for "mild cognitive impairment" genes and then selecting the pair with the lowest number of those genes could give a rate of successful guessing well above random. So this study, alone, would not be likely to allow you to make a prediction that's much better than random, and resumes are going to be hard to beat period (though with the resumes and the genetic data, you might be able to make better predictions). And if Zhao and his team do find a gene that correlates with increased intelligence, that will be big news (though I would not give good odds on that actually happening). In any case, I agree that IQ is complex, but I don't think that either the "intelligence gene" hypothesis or Dickens and Flynn's hypothesis is the whole story, because identical twins raised apart still have a significant and fairly high correlation in IQ scores [2]. [1] http://duepublico.uni-duisburg-essen.de/servlets/DerivateServlet/Derivate-28058/Oades2005analysis.pdf http://duepublico.uni-duisburg-essen.de/servlets/DerivateSer... [2]http://onlinelibrary.wiley.com/doi/10.1111/j.2044-8295.1966.tb01014.x/abstract http://onlinelibrary.wiley.com/doi/10.1111/j.2044-8295.1966.... (unfortunately paywalled, though I'm sure someone will accidentally drop it if you need it).
rsheridan6 14y ago
(4) is very plausible even if the study fails to find genetic variants associated with intelligence. Assuming intelligence is heritable at all, you don't need to understand the underlying genetic architecture to breed for it any more than neolithic farmers needed to understand the genetics of docile sheep to breed them. This has been known for a long time, but there has been no large scale, serious attempt to implement a eugenics program breeding for intelligence. I don't see how this research would change anything, successful or not, other than by getting the subject some attention in the news.

by bthomas 14y ago

Scaevolus 14y ago

Why do you think discovering genetic components to intelligence is extremely unlikely? I expect it to be easier to find genetic components for intelligence than for other polygenetic traits like height -- the brain has fewer components influencing development.

bthomas 14y ago
> the brain has fewer components influencing development What do you mean by that? I'd argue the opposite. In a sense, we already have discovered genes for high height with Marfan Syndrome. I think IQ is orders of magnitude more complex. Awesome conversation but probably hard in an online forum, so I'll just cite my favorite example: Dickens and Flynn [1] I absolutely think this study has a chance of revealing some signal, my skepticism is just that it will have any actionable conclusions. Here's a litmus test: what results would allow you to choose from a group of 10 men and 10 women, based only on genotype, which pair is most likely to have a high IQ child. And how would your power compare to just reading all their resumes? [1] http://en.wikipedia.org/wiki/Heritability_of_IQ#Dickens_and_Flynn_model http://en.wikipedia.org/wiki/Heritability_of_IQ#Dickens_and_...
- Scaevolus 14y ago
  Barring specific diseases like Marfan Syndrome, we can't yet explain most of height -- "Height is a highly heritable trait, with estimates of heritability as high as 90%. Recent genome-wide association studies of height have discovered over 180 common variants associated with height. These variants have small effect sizes and collectively explain approximately 10% of the heritability." [1] Growth involves many systems with many types of cells interacting -- digestive, skeletal, endocrine, etc. There's a lot of complexity when you have so many systems involved. By comparison, there are relatively few distinct types of cells in the brain, so the effects of different alleles should be more apparent. In particular, genes controlling sensitivity to and production of neurotransmitters directly impact brain function. If they find many specific mutations that cause better performance on specific tasks (working memory, pattern matching, verbal ability, ...) it should be possible to predict high IQs. I don't expect 100% accurate predictions, but to not find _anything_ would be surprising. Maybe that's your point-- if they only get 10% heritability, it's not really enough for breeding decisions. [1] http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002439 http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fj...
- JoshuaDavid 14y ago
  I recognize that your litmus test was rhetorical, but it's a question that does seem answerable, so I'll try to answer it. I would imagine there are a number of genes that are associated with lowered intelligence. For example, one might start by looking at genes such as the Dopamine D4 receptor gene, for which the 7-repeat allele seems to be a pretty good indicator of ADHD [1][2], which in turn is associated with significantly lower IQ. Of course, chains of correlation don't always work, and that is why it would be important to have tens or preferably hundreds to thousand of things like that to look for. While it would be nice and easy to look for an "intelligence gene", that may not be feasible. However, my guess is that looking for "mild cognitive impairment" genes and then selecting the pair with the lowest number of those genes could give a rate of successful guessing well above random. So this study, alone, would not be likely to allow you to make a prediction that's much better than random, and resumes are going to be hard to beat period (though with the resumes and the genetic data, you might be able to make better predictions). And if Zhao and his team do find a gene that correlates with increased intelligence, that will be big news (though I would not give good odds on that actually happening). In any case, I agree that IQ is complex, but I don't think that either the "intelligence gene" hypothesis or Dickens and Flynn's hypothesis is the whole story, because identical twins raised apart still have a significant and fairly high correlation in IQ scores [2]. [1] http://duepublico.uni-duisburg-essen.de/servlets/DerivateServlet/Derivate-28058/Oades2005analysis.pdf http://duepublico.uni-duisburg-essen.de/servlets/DerivateSer... [2]http://onlinelibrary.wiley.com/doi/10.1111/j.2044-8295.1966.tb01014.x/abstract http://onlinelibrary.wiley.com/doi/10.1111/j.2044-8295.1966.... (unfortunately paywalled, though I'm sure someone will accidentally drop it if you need it).

rsheridan6 14y ago

(4) is very plausible even if the study fails to find genetic variants associated with intelligence. Assuming intelligence is heritable at all, you don't need to understand the underlying genetic architecture to breed for it any more than neolithic farmers needed to understand the genetics of docile sheep to breed them. This has been known for a long time, but there has been no large scale, serious attempt to implement a eugenics program breeding for intelligence. I don't see how this research would change anything, successful or not, other than by getting the subject some attention in the news.