RSS
The search for genetic variants that contribute to higher intelligence has been slow, perhaps unsurprisingly. After all, your brain is, arguably, the most complicated thing in the known universe. Here’s a big new paper that might mark a milestone in this research.
Significance
We identify several common genetic variants associated with cognitive performance using a two-stage approach: we conduct a genome-wide association study of educational attainment to generate a set of candidates, and then we estimate the association of these variants with cognitive performance. In older Americans, we find that these variants are jointly associated with cognitive health. Bioinformatics analyses implicate a set of genes that is associated with a particular neurotransmitter pathway involved in synaptic plasticity, the main cellular mechanism for learning and memory. In addition to the substantive contribution, this work also serves to show a proxy-phenotype approach to discovering common genetic variants that is likely to be useful for many phenotypes of interest to social scientists (such as personality traits).
Abstract
We identify common genetic variants associated with cognitive performance using a two-stage approach, which we call the proxy-phenotype method. First, we conduct a genome-wide association study of educational attainment in a large sample (n = 106,736), which produces a set of 69 education-associated SNPs. Second, using independent samples (n = 24,189), we measure the association of these education-associated SNPs with cognitive performance. Three SNPs (rs1487441, rs7923609, and rs2721173) are significantly associated with cognitive performance after correction for multiple hypothesis testing. In an independent sample of older Americans (n = 8,652), we also show that a polygenic score derived from the education-associated SNPs is associated with memory and absence of dementia. Convergent evidence from a set of bioinformatics analyses implicates four specific genes (KNCMA1, NRXN1, POU2F3, and SCRT). All of these genes are associated with a particular neurotransmitter pathway involved in synaptic plasticity, the main cellular mechanism for learning and memory.
Lots of senior big names, such as Plomin, Deary, Pinker, Chabris, and Conley, feature as co-authors.
English
Arabic
Chinese (Simplified)
Dutch
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Persian
Polish
Portuguese
Russian
Spanish
Swedish
Turkish
Vietnamese
Until we see the P values, it’s suspect.
Association studies tend to be garbage (non-replicable). Of course when you look at 10,000 different factors (such as SNP’s), a few will be statistically significant by pure chance.
There were 69 SNPs tested in that stage so with Bonferroni correction the threshold is 0.05/69 = 7.2e-4 and only the first three are significant. (my interpretation, I am not a professional in this field)
Association studies tend to be garbage (non-replicable). Of course when you look at 10,000 different factors (such as SNP's), a few will be statistically significant by pure chance.Replies: @Steve Sailer, @res
The SNPs they found with the big sample also worked on a second, independent sample.
The whole paper is free online, so why don’t you read it?
http://www.pnas.org/content/early/2014/09/05/1404623111.full.pdf+html
Is this similar to the massive project being done by an institute in China? I’m referring to the one where they solicited DNA samples from very bright people all over the world . I would have sent in a sample but am sadly not bright enough to qualify.
Sheeeeeeeeeeeet.
http://www.amren.com/news/2014/09/political-cartoon-contest-winners/
Until we see the P values, it’s suspect.
As the abstract says, the p values are significant even after correction for multiple hypotheses. The paper says the correction is the Bonferroni correction, which is as conservative as it gets. The corrected p values are all about .02.
If there’s a problem with the study, it’s somewhere else.
The effect size of each of the genes is pretty minuscule, though:
head in the clouds
http://www.straightdope.com/columns/read/1263/can-a-cloud-weigh-as-much-as-a-747
http://www.straightdope.com/columns/read/1273/how-are-shrunken-heads-made
http://www.straightdope.com/columns/read/3177/can-a-girl-lose-her-virginity-riding-a-roller-coaster
James Thompson discusses our grim future:
http://www.drjamesthompson.blogspot.co.uk/2014/09/the-united-states-of-mexico.html#comment-form
https://www.ets.org/Media/Education_Topics/pdf/AmericasPerfectStorm.pdf
The author list is a who’s who of geneticists interested in IQ. But what’s Pinker doing in there? That guy just can’t stay out of trouble.
I wonder how all those fine geneticists who signed that letter to The NY Times protesting Nicholas Wade’s “misrepresentation” of their science are going to feel when they read this result.
There’s nothing cross-racial about this study, of course, but how much of a leap can it be to such a further result?
I spy a bead of sweat on their brows.
As the abstract says, the p values are significant even after correction for multiple hypotheses. The paper says the correction is the Bonferroni correction, which is as conservative as it gets. The corrected p values are all about .02.
If there's a problem with the study, it's somewhere else.
The effect size of each of the genes is pretty minuscule, though:Replies: @BehindTheLines
This roughly correlates with Steve Hsu’s estimate. He estimated that 10000 alleles effect IQ, with the good alleles having 90% frequency and the bad ones having 10% frequency. Assume a Poisson distribution and you get each person having 1000 bad alleles with a standard deviation of about 30. So, each bad allele would subtract 15/30 = 0.5 IQ points. Pretty close to the 0.3 of this paper.
You read the paper a bit wrong I think. Or maybe I did, so please correct me if I'm wrong, but they found 3 genes which have a significant effect on IQ equal to .036, -.034, and -.034 of an IQ SD, which is confusingly how they state their results. So that is +0.5 IQ points for one, -0.5 IQ points for two of the variants.
Also, I am reading this right in that all three genes have a very close to 50% general population frequency and that they excluded all <1% variants? Otherwise they would have picked up on the Ashkenazi rare variants that a worth a lot more than half an IQ point.
It is also somewhat arbitrary to call one or the other a variants when they are so close to 50% frequency, so don't read anything into the authors choice of a positive result for one gene and a negative result for two more.Replies: @res
Association studies tend to be garbage (non-replicable). Of course when you look at 10,000 different factors (such as SNP's), a few will be statistically significant by pure chance.Replies: @Steve Sailer, @res
P values are in Table S4 of http://www.pnas.org/content/suppl/2014/09/06/1404623111.DCSupplemental/pnas.1404623111.sapp.pdf
There were 69 SNPs tested in that stage so with Bonferroni correction the threshold is 0.05/69 = 7.2e-4 and only the first three are significant. (my interpretation, I am not a professional in this field)
OT:
For anyone still unsure of the heritability of IQ– a candid interaction with those in the other half.
A local news crew is interviewing neighbors of a resident whose unleashed dog retrieved a human skull overnight. The neighborhood is across the street from the Austin Airport and is a mix of mobile homes and manufactured housing– not homeless folks.
The smaller sample used by the authors are Americans only. The larger sample is a mix of Americans, Australians, English, and Swedish. Most of the studies include people of all ethnicity in those countries.
BGI is looking at both Chinese and Americans, and in particular is looking at math ability genes.
Also, wow, I have never quite seen a paper with more authors and more acknowledgements.
Here is just the “A to D” of the acknowledgements page:
Safety in numbers, huh? Seems that there is going to be a lot more to come!
It may be of interest to some that SNP rs7923609 is included in 23and me’s genotyping service. You can access it by clicking on your user name in the upper right corner, selecting browse raw data, and entering the number in the SNP field.
Not that one gene out of the thousands that affect IQ really tells you anything you didn’t already know about how smart you are but it can be interesting just to look these things up nonetheless.
http://www.haematologica.org/content/haematol/early/2011/05/12/haematol.2011.042077.full.pdf
Seems to have a huge effect on blood clotting, much bigger than its IQ effect.Replies: @Steve Sailer, @candid_observer
Pinker’s sole contribution was reading the manuscript and providing the comments: “Steven Pinker … critically reviewed and edited the manuscript.” I.e., he’s got no business to be in the author list and his name appearing there is pretty unethical.
“So, each bad allele would subtract 15/30 = 0.5 IQ points. Pretty close to the 0.3 of this paper.”
You read the paper a bit wrong I think. Or maybe I did, so please correct me if I’m wrong, but they found 3 genes which have a significant effect on IQ equal to .036, -.034, and -.034 of an IQ SD, which is confusingly how they state their results. So that is +0.5 IQ points for one, -0.5 IQ points for two of the variants.
Also, I am reading this right in that all three genes have a very close to 50% general population frequency and that they excluded all <1% variants? Otherwise they would have picked up on the Ashkenazi rare variants that a worth a lot more than half an IQ point.
It is also somewhat arbitrary to call one or the other a variants when they are so close to 50% frequency, so don't read anything into the authors choice of a positive result for one gene and a negative result for two more.
BGI is looking at both Chinese and Americans, and in particular is looking at math ability genes.
Also, wow, I have never quite seen a paper with more authors and more acknowledgements.Here is just the "A to D" of the acknowledgements page:Safety in numbers, huh? Seems that there is going to be a lot more to come!Replies: @a very knowing American
In Classical China, there was a method of execution called something like “A Thousand Bites” (sorry, not sure exactly what it was, I’m away from references at the moment). The idea was that each of a large group (a village, say, or a guild) would bite a miscreant. No one bite was fatal, so the governing Mandarin could be told that no one individual was a murderer, but the cumulative effect was to kill the victim. What we may be seeing here is the death of the No IQ/Genes/Race nexus by A Thousand Bites.
"The rule of the [gold beater's guild] was that no employer could have more than one apprentice at a time, but one member of the craft represented to the magistrate that, if he were allowed to take on a number of apprentices, the work would be expedited. He received permission to do so and engaged a great many apprentices. This ... conduct infuriated the craft. The word was passed around that "biting to death is not a capital offense," apparently on the gruesome theory that no one morsel is fatal, and the cartel-buster was soon dead from the fiendish efforts of 123 of his fellows. None of the guild members was allowed to leave the shop until his teeth and gums attested to what, in more delicate settings, might be called his "professional ethics." However "the man who took the first bite was, it turns out, discovered and executed."
Mancur Olson, The Rise and Decline of Nations, p. 149
So, again if I am reading this right, we can now predict IQ from genes, using ONLY the information in this paper about three genes, but not very well. Those of us who roll three A’s have an average IQ of 100.75, those with three B’s 99.25.
If you combine all less significant variants, maybe you’d triple that range, but very few people would have the extreme range of prediction, and even then only a predicted 102.25 or 97.75
“Author contributions: D.J.B., D. Cesarini, and P.D.K. designed research; C.A.R., T.E., G.D., T.H.P., P.T., B.B., V.E., A.D.J., J.J.L., C.d.L., R.E.M., S.E.M., M.B.M., O.R., S.J.v.d.L., A.A.E.V., N.A., D. Conley, J.D., R.F., L.F., C.H., C.I.-V., J.K., D.C.L., P.K.E.M., G.M., D.P., M.T., M.E.W., M.J., P.M.V., and D. Cesarini analyzed data; C.A.R., T.E., P.T., C.F.C., D.L., D.J.B., D. Cesarini, and P.D.K. wrote the paper; C.F.C., C.M.v.D., E.L.G., W.G.I., V.J., D.L., P.L., N.G.M., M.M., N.L.P., S.P., D.P., J.M.S., H.T., F.C.V., M.J.W., G.D.S., I.J.D., M.J., and R.P. performed data collection; J.J.L., M.B.M., C.M.v.D., N.K.H., P.K.E.M., D.J.P., B.H.S., J.M.S., H.T., N.J.T., M.J.W., I.J.D., and M.J. performed phenotyping; and G.D., M.B.M., C.M.v.D., C.H., V.J., D.C.L., P.K.E.M., N.G.M., D.J.P., F.R., N.J.T., and A.G.U. performed genotyping.”
So S.P. and some other Big Initials are down as “performed data collection.”
We’ve really forgotten our heritage if people are using random anecdotes from Chinese history when there is a perfect Classical example.
The death of Caeser was by stabbing, one stab per Senate conspirator so all would be equally culpable.
Also, I really doubt the ancient Chinese killed each other by mass biting. That’s disgusting, even by barbaric ancient standards.
You read the paper a bit wrong I think. Or maybe I did, so please correct me if I'm wrong, but they found 3 genes which have a significant effect on IQ equal to .036, -.034, and -.034 of an IQ SD, which is confusingly how they state their results. So that is +0.5 IQ points for one, -0.5 IQ points for two of the variants.
Also, I am reading this right in that all three genes have a very close to 50% general population frequency and that they excluded all <1% variants? Otherwise they would have picked up on the Ashkenazi rare variants that a worth a lot more than half an IQ point.
It is also somewhat arbitrary to call one or the other a variants when they are so close to 50% frequency, so don't read anything into the authors choice of a positive result for one gene and a negative result for two more.Replies: @res
I think the 0.3 IQ points number comes from the corrected coefficients in Table S7 of the supplemental material.
And they say 0.036, etc, not 3. They are talking about standard deviations. So to translate to IQ, you need to multiply the IQ SD of 15 by .036, which is 0.54.
Don't worry if you didn't notice this, I had to spend a fair amount of time to figure this out myself. Despite being the big headline result, they don't ever say it directly.Replies: @res, @res
S7 does not say anything about IQ points directly.
And they say 0.036, etc, not 3. They are talking about standard deviations. So to translate to IQ, you need to multiply the IQ SD of 15 by .036, which is 0.54.
Don’t worry if you didn’t notice this, I had to spend a fair amount of time to figure this out myself. Despite being the big headline result, they don’t ever say it directly.
I assume you meant well by your next to last sentence, but it is easy to read that as patronizing.
Perhaps I should apologize for not including the intermediate steps which would have made my reasoning in the earlier post more clear.Replies: @Lot
Not that one gene out of the thousands that affect IQ really tells you anything you didn't already know about how smart you are but it can be interesting just to look these things up nonetheless.Replies: @Lot
Here’s the most interesting other result about that gene:
http://www.haematologica.org/content/haematol/early/2011/05/12/haematol.2011.042077.full.pdf
Seems to have a huge effect on blood clotting, much bigger than its IQ effect.
It seems to me that the beauty of the result is that it allows others to independently confirm the effect of these SNPs/alleles on a dataset of their choosing. Those sorts of tests should be pretty basic and simple, and repeated confirmations should be hard to explain in any other way than that these SNPs/alleles do indeed have an impact on cognitive performance.
http://www.haematologica.org/content/haematol/early/2011/05/12/haematol.2011.042077.full.pdf
Seems to have a huge effect on blood clotting, much bigger than its IQ effect.Replies: @Steve Sailer, @candid_observer
Blood clotting is of course a factor in strokes in the brain.
http://en.wikipedia.org/wiki/Stoning
“No individual among the group can be identified as the one who kills the subject.”
And they say 0.036, etc, not 3. They are talking about standard deviations. So to translate to IQ, you need to multiply the IQ SD of 15 by .036, which is 0.54.
Don't worry if you didn't notice this, I had to spend a fair amount of time to figure this out myself. Despite being the big headline result, they don't ever say it directly.Replies: @res, @res
Table S7 allows you to infer the IQ points for the corrected coefficients just as you did in your post for the uncorrected coefficient. The corrected coefficient of about 0.023 corresponds to 0.345 IQ points. The other two SNPS are closer to 0.3 IQ points. That was why I wrote what I did.
I think you meant well by your next to last sentence, but it is easy to read that as patronizing.
OK, further research confirms that the frequency listed for these genes is indeed all near one half, at least for the mostly-white study sample. The higher-IQ G rs7923609 frequency is about 52% for euros, 31% of Asians and 37% for Africans.
The other big result of the study is having the high-IQ genes reduces elderly cognitive decline.
Great study, but the results are not stated very clearly!
And they say 0.036, etc, not 3. They are talking about standard deviations. So to translate to IQ, you need to multiply the IQ SD of 15 by .036, which is 0.54.
Don't worry if you didn't notice this, I had to spend a fair amount of time to figure this out myself. Despite being the big headline result, they don't ever say it directly.Replies: @res, @res
Table S7 allows one to infer IQ points for the corrected coefficients just as you did for the uncorrected coefficients. From Table S7 we have the corrected coefficient 0.023 * 15 = 0.345 IQ points (i.e. about 0.3, especially since the others SNPs are close to 0.02 corrected).
I assume you meant well by your next to last sentence, but it is easy to read that as patronizing.
Perhaps I should apologize for not including the intermediate steps which would have made my reasoning in the earlier post more clear.
Personally, I have no idea if the corrections should be applied or not to improve the data. The authors, at least, don't include the corrected table in their paper, just the appendix, so they seemingly preferred the uncorrected figure.
If I'm wrong, and these are true "corrections" to the effect size and not mere robustness checks (1) Why not put them in the main paper? (2) Why did they do three separate adjustments and seemingly not combine or state a preference for one of the three?
I assume you meant well by your next to last sentence, but it is easy to read that as patronizing.
Perhaps I should apologize for not including the intermediate steps which would have made my reasoning in the earlier post more clear.Replies: @Lot
The “uncorrected” data show ~0.5 IQ points for all three genes, and the authors offer three possible corrections to the raw data. They don’t consistently show 0.3 IQ points, just one of the three. The other two are far less than 0.3.
Personally, I have no idea if the corrections should be applied or not to improve the data. The authors, at least, don’t include the corrected table in their paper, just the appendix, so they seemingly preferred the uncorrected figure.
If I’m wrong, and these are true “corrections” to the effect size and not mere robustness checks (1) Why not put them in the main paper? (2) Why did they do three separate adjustments and seemingly not combine or state a preference for one of the three?
So, esoteric jargon aside…does this give HBD deniers a merciless, brutal, bloody knockout blow…or not?
My take on it is that it provides a more convincing case that cognitive performance has a genetic basis within populations -- though not yet between populations.
One might say that the previous result of Visscher et al
http://www.nature.com/mp/journal/v16/n10/full/mp201185a.html
already demonstrated this, and indeed set a minimum lower bound of 51% for the genetic basis of fluid g. But the techniques employed were conceptually quite complex, and the argument therefore less broadly convincing. Moreover, there was no real understanding advanced of which sorts of alleles might affect such performance, or what the mechanism might be.
But everybody understands the concept of an individual SNP/allele. It should be easy to verify the result -- indeed, far easier than to verify it in this study itself. Presumably, because further confirmatory studies can focus just on these three SNPs, if they use truly independent datasets, the methods won't require the same high level of multiple hypothesis correction as in this study. And of course the effect of the three SNPs can be investigated regarding their mechanism of action.
If the result holds up, it is certainly the death of the idea that all differences in cognitive performance are culturally based, which, amazingly, a lot of people do seem to believe, or at minimum hold out hope might be so. And it becomes far more difficult to crimestop the further thought that difference races might differ on these very traits.
China no rotherham
http://alternative-right.blogspot.com/2014/09/the-acceptable-enemy.html#more
Steve, you missed your favorite author: Edward L. Glaeser.
The death of Caeser was by stabbing, one stab per Senate conspirator so all would be equally culpable.
Also, I really doubt the ancient Chinese killed each other by mass biting. That's disgusting, even by barbaric ancient standards.Replies: @Laban
It’s “death by a thousand cuts” not bites afaik
http://en.wikipedia.org/wiki/Slow_slicing
Note that this is the consortium that previously found education correlations. They were also the ones who previously pointed out that all existing IQ-SNP results were p-value induced crap.
Note that their education results survived some pretty hostile falsification attempts (including large-scale independent sample replication). I´d expect the same this time around.
“Lots of senior big names, such as Plomin, Deary, Pinker, Chabris, and Conley, feature as co-authors.”
The usual question in such cases is “which of them has even read the paper?” The answer would seem to be “Well, at least Pinker has.”
I’m definitely a non-expert. If these are the only significant SNP’s they found, how do we get to a population IQ SD of 15 with three effects on the order of 0.3?
Maybe the actual researchers wanted his nane included to add credibility and provide protection.
It’s hard to tell. That role can mean anything from “did absolutely nothing” (usually means this, I believe) to “made extensive excellent comments leading to a major improvement in the paper.” The latter arguably deserves authorship.
It is worthwhile reading this paper:
http://arxiv.org/pdf/1408.3421v1.pdf
It sets out a very good argument for why we can ignore non-additive (gene x gene) contributions to IQ. They are pretty much not heritable because of crossover during meiosis.
Intelligence-enhancing-allele frequencies for dozens of ethnic groups have already been calculated by Piffer (“Factor Analysis of Population Allele Frequencies…” Mankind Quarterly, vol. 54, 2013). There is about a 1 1/2 standard deviation advantage for Europeans vs. sub-Saharan Africans (American blacks have a similar allele frequency, despite white admixture). Han Chinese have almost a one standard deviation advantage over Europeans. The Finns outscore the French. Pygmies and Bushmen score lowest.
@shamu
No. Evidence is not welcome, or listened to by the Cathedral/orthodoxy at this time.
It it the first trickle through the dam in the sense that it now might now be OK for more researchers to do research on related topics, even in an attempt at disproval.
The fact that Steve Hsu was hired in his current position to me means that there are friends of the truth in large research institutions. They are hamstrung by enormous political problems due to the prevailing political culture at these institutions.
To remain relevant in the frontiers of scientific discovery (which is their reason for the existence of the modern research university) administrators have to find a way to make these inquiries not toxic to researchers careers and and grant proposals. The large number of participants (a large school of researchers) and the addition of a few whales like Pinker aid the cause.
If it does not work, the Chinese will be happy to take the bulk of the work from the West over the next 20 years.
So, esoteric jargon aside…does this give HBD deniers a merciless, brutal, bloody knockout blow…or not?
My take on it is that it provides a more convincing case that cognitive performance has a genetic basis within populations — though not yet between populations.
One might say that the previous result of Visscher et al
http://www.nature.com/mp/journal/v16/n10/full/mp201185a.html
already demonstrated this, and indeed set a minimum lower bound of 51% for the genetic basis of fluid g. But the techniques employed were conceptually quite complex, and the argument therefore less broadly convincing. Moreover, there was no real understanding advanced of which sorts of alleles might affect such performance, or what the mechanism might be.
But everybody understands the concept of an individual SNP/allele. It should be easy to verify the result — indeed, far easier than to verify it in this study itself. Presumably, because further confirmatory studies can focus just on these three SNPs, if they use truly independent datasets, the methods won’t require the same high level of multiple hypothesis correction as in this study. And of course the effect of the three SNPs can be investigated regarding their mechanism of action.
If the result holds up, it is certainly the death of the idea that all differences in cognitive performance are culturally based, which, amazingly, a lot of people do seem to believe, or at minimum hold out hope might be so. And it becomes far more difficult to crimestop the further thought that difference races might differ on these very traits.
http://www.haematologica.org/content/haematol/early/2011/05/12/haematol.2011.042077.full.pdf
Seems to have a huge effect on blood clotting, much bigger than its IQ effect.Replies: @Steve Sailer, @candid_observer
Seems to have a huge effect on blood clotting, much bigger than its IQ effect.
It certainly is plausible that this allele actually may have a tradeoff effect in terms of survival/reproduction, so that the additional boost to survival/reproduction via an increase in cognitive performance is offset by a decrease in survival due to strokes.
If the allele can be found in all races, and at close to 50% frequency, then it is probably of very old origin, and the fact that it has not yet gone to fixation, or near fixation, despite presumably strong selection for cognitive performance as a trait, suggests that there is a tradeoff involved. That’s probably why it is, apparently, one of SNPs of greatest effect on cognitive performance. Most likely, alleles with a purely positive impact on survival, exclusively via a boost to cognitive performance, have already gone to fixation unless they are relatively new or of very small effect.
Here’s what I was half-remembering about biting people to death:
“The rule of the [gold beater’s guild] was that no employer could have more than one apprentice at a time, but one member of the craft represented to the magistrate that, if he were allowed to take on a number of apprentices, the work would be expedited. He received permission to do so and engaged a great many apprentices. This … conduct infuriated the craft. The word was passed around that “biting to death is not a capital offense,” apparently on the gruesome theory that no one morsel is fatal, and the cartel-buster was soon dead from the fiendish efforts of 123 of his fellows. None of the guild members was allowed to leave the shop until his teeth and gums attested to what, in more delicate settings, might be called his “professional ethics.” However “the man who took the first bite was, it turns out, discovered and executed.”
Mancur Olson, The Rise and Decline of Nations, p. 149
You mean 3 out of thousands replicated in two samples?
Also where are all the other previous alleles that were associated in other studies? There was another 3 that were replicated twice, with big populations like this but didn’t show up in this one.
Wheres the rest?
Where is the molecular test on said allele to show what it even does? Do they know? Or is that also just based off an association and an assumption?
Also those other 3 I mentioned didn’t show up in another 2 big association studies in mostly Europeans like this. These 3 also didn’t show up in any of those either.
The first was in 2012 that found HGMA2 correlation n50 000 but none of these then another of Australian Europeans with about n9000 that found none of these too.
4 studies with enough sample size did not find these 3.
More nonsense by the old genetic determinism brigade by the looks of it.
The 4 studies is including the previous two that Hsu posted on his blog and the two I mentioned above.
Whether these 3 SNPs would show up in a given association study depends entirely on the method the study used to come up with its list of candidates. Because of the problems with correcting for multiple hypotheses, it's easy for the list generated to be composed entirely of false positives while excluding all true positives. To relax the corrections so that true positives turn up even when their effect is as small as these SNPs, the list might have to be so stupendously long and dominated by false positives that it would be useless.
Again, it should be a different ball game going forward, with the SNPs in hand. It should be pretty easy to confirm -- or discomfirm -- this result.
The first was in 2012 that found HGMA2 correlation n50 000 but none of these then another of Australian Europeans with about n9000 that found none of these too.
4 studies with enough sample size did not find these 3.
More nonsense by the old genetic determinism brigade by the looks of it.Replies: @Hoho, @candid_observer
Excuse me a clarification:
The 4 studies is including the previous two that Hsu posted on his blog and the two I mentioned above.
The first was in 2012 that found HGMA2 correlation n50 000 but none of these then another of Australian Europeans with about n9000 that found none of these too.
4 studies with enough sample size did not find these 3.
More nonsense by the old genetic determinism brigade by the looks of it.Replies: @Hoho, @candid_observer
Also those other 3 I mentioned didn’t show up in another 2 big association studies in mostly Europeans like this. These 3 also didn’t show up in any of those either.
Whether these 3 SNPs would show up in a given association study depends entirely on the method the study used to come up with its list of candidates. Because of the problems with correcting for multiple hypotheses, it’s easy for the list generated to be composed entirely of false positives while excluding all true positives. To relax the corrections so that true positives turn up even when their effect is as small as these SNPs, the list might have to be so stupendously long and dominated by false positives that it would be useless.
Again, it should be a different ball game going forward, with the SNPs in hand. It should be pretty easy to confirm — or discomfirm — this result.
What milestone?
Basically its another study in which they haven’t found any genes for IQ, only associated 3 out of thousands that could maybe be involved in causing IQ differences between people.
The only milestone is going to come when the entire dogma of this neo Darwinian nonsense comes to an end.
You're only saying that because you can't understand things like:
On the genetic architecture of intelligence and other quantitative traits
“The SNPs they found with the big sample also worked on a second, independent sample.”
More accurate would be only 3 out 63 associations worked and only 63 out of thousands even showed a sign in the first place. The 3 didn’t show up in many other studies too.
That would be accurate.
This is interesting and encouraging work – necessary so that further work can be done – but it doesn’t actually speak to the individual variations that concern us the most. The trait known as ‘time preference’ is the single most significant example.
All of the emotional preferences and inclinations that statistics tell us are grossly unequal among various groups of people, and which are even more critical to intellectual function and achievement than IQ, are not spoken to by the study.
I am not willing to speculate about precisely what biological factors go into time preference, but the safe money is that it’s not an accumulation of tiny, additive factors. I rather doubt there’s a single gene – can you distinguish Nabokov from Wolfe by looking at a single word from their respective works? – but probably a small collection of genes that change some vital structures. A series of paragraphs, perhaps together, perhaps scattered through the collected works.
I don’t get the appeal of this whole “hunt for the IQ genes”. It’s not like we are living in 1985 and know nothing on the subject. By now, the preponderance of evidence says that there must be many thousands, perhaps even every other, human genes that contribute to IQ. Some clueless HBD believers have bought, however, into the whole [stupid] premise of “if you can’t show the gene(s), you can’t be sure it is genetics-based”. Well, no! There is exactly zero doubt about populational differences in IQ, and there is very, very little doubt that hereditary contribution is very large, most likely accounting for the majority of variation. Putting together a list of genes won’t affect any of the conclusions above a jot!
I hate to say this because I presume you’re one of the good guys, but your perspective is naïve. There’s already overwhelming scientific proof that differences in cognitive performance have genetic bases both among persons and among “populations”. There isn’t even a competing theory!
It’s not a matter of proof; it’s a matter of religious faith and intimidation.
Thank you for that fascinating story. Olson has a lot of interesting things to say.
Basically its another study in which they haven't found any genes for IQ, only associated 3 out of thousands that could maybe be involved in causing IQ differences between people.
The only milestone is going to come when the entire dogma of this neo Darwinian nonsense comes to an end.Replies: @The most deplorable one
Ha ha.
You’re only saying that because you can’t understand things like:
On the genetic architecture of intelligence and other quantitative traits