Common variant for "IQ gene"?

By Razib Khan | April 15, 2012 11:11 pm

A few people have forwarded me this paper, Identification of common variants associated with human hippocampal and intracranial volumes:

…Whereas many brain imaging phenotypes are highly heritable…identifying and replicating genetic influences has been difficult, as small effects and the high costs of magnetic resonance imaging (MRI) have led to underpowered studies. Here we report genome-wide association meta-analyses and replication for mean bilateral hippocampal, total brain and intracranial volumes from a large multinational consortium. The intergenic variant rs7294919 was associated with hippocampal volume (12q24.22; N = 21,151; P = 6.70 × 10−16) and the expression levels of the positional candidate gene TESC in brain tissue. Additionally, rs10784502, located within HMGA2, was associated with intracranial volume (12q14.3; N = 15,782; P = 1.12 × 10−12). We also identified a suggestive association with total brain volume at rs10494373 within DDR2 (1q23.3; N = 6,500; P = 5.81 × 10−7).

Look at the sample sizes. Beware of behavior genomics with small sample sizes. Paul Thompson, one of the many authors of this paper, is giving media interviews. To me that’s a good sign, as he’s a very smart guy. He has some confidence in this study. Here’s the section which is resulting in the forwards:

… In addition, the C allele of rs10784502 is associated, on average, with 9,006.7 mm3 larger intracranial volume, or 0.58% of intracranial volume per risk allele and is weakly associated with increased general intelligence by approximately 1.29 IQ points per allele.

I’m a homozygote for the T allele for what it’s worth. But that’s not surprising. Look at the population distribution of the C allele from the HapMap:

I think CD-CV is played out when it comes to IQ or height. Note that the gene in which this SNP is located, HMGA2, is also implicated in height variation. In fact, aside from pathological mutations, this locus may have the largest effect height and IQ polymorphisms in replicable studies (genetic and genomic, linkage and GWAS, IQ studies have a history of not being replicated, probably because of low power).

Update: See Luke Justins’ comment. My post above was rooted in a mild misipression due to read-post-and-dash. My bad. So we’ll see if the ‘endophenotype’ pans out.

MORE ABOUT: Genetics, Genomics, IQ

Comments (15)

  1. I’d be less confident about the IQ association than the cranial volume association. The sample size wasn’t massive (N=~1600) and the association was relatively weak (p = 0.0073) – we know that studies larger than this have not been well powered to reliably detect associations to IQ measures. Also, Davies et al [1] don’t list this association in their supp tables (which contained all associations with p < 10^-4 in their scan), and their sample size (N=3511) would have had a 95% power to detect an association with that effect size at p < 10^-4 (beta = 1.29/15). Differences in the way IQ is measured/different SNP chips/etc etc could account for this, but I'd be wary of saying this is a bona fide IQ association.

    Not really sure why they didn't just look up this association in the whole meta-analysis data, given the overlap in authors.


  2. ackbark

    Can this be a remnant of archaic populations having larger brains?

    “We also identified a suggestive association with total brain volume at rs10494373 within DDR2. . .”

    But surely you can just buy more RAM. . .

  3. “Look at the sample sizes. Beware of behavior genomics with small sample sizes.”

    I’m a little confused by this comment. These samples are huge for an MRI-genotype association study, which often have samples sizes in the hundreds or dozens or less, which the highly significant p values reflect. Is the implication of the comment that past studies with smaller sample sizes shouldn’t have be taken as seriously as this one, or that this study in particular is too small?

    Also, an MRI-genotype association study has the virtue, not common to other kinds of studies, of involving very well defined and basically indisputable data points, rather than, for example, personality surveys or IQ tests, as the core measured elements. This greatly reduces inherent systemic error associated with the measurement process, leaving only sampling error which the large sample size mutes.

    I also don’t see any particularly good reason to distrust at first glance a weak association of cranial volume and IQ, as an obvious mechanism is available to relate the two, and IQ while hardly a perfect operational test is one of the best defined and most accurately operationalized metrics in psychology (certainly far more accurate than most personality measures or even than mos psychiatric diagnoseses). The claimed total effect size of the entire set of allelles appears to be less than a third of a standard deviation in any case, so it isn’t as if someone is proposing a grant overall theory of intellligence here.

    The biggest concern in my view in this study would be that a merely ancestry informative marker with neutral effect on intelligence is picking up of intelligence differences attributable to histories of immigration instead (i.e. many U.S. Asian immigrants were able to immigrate due to their advanced educational degrees making them unrepresentative of Asians generally, while the African slave trade was not selective in that was and has been exacerabated by strong non-hereditary effects on IQ among the many low income African-Americans today; so any marker that is more common in Asians than Africans, even if it has no causal connection to IQ relevant brain function, is going to be associated statistically with IQ in any U.S. drawn sample).

    To test that possibility, you’d really want to see if the cranial size and IQ correlations held up just as strongly within demographic subgroups as they do across demographic subgroups. Indeed, given environmental impacts due to poverty, the intragroup assocation should be stronger than the intergroup association, and the association should be stronger among middle class or better subjects than among poor subjects (which the data might contain some proxy to determine).

  4. I’ve to admit that I still haven’t found the time to read the publication in detail, but do you know if they corrected the intracranial volume for the corresponding heights?

    I think it’s generally acknowledged that height and intracranial volume correlate (Dawkins has a nice take on this in «The Ancestor’s Tale». So given the relatively small differences in volume it seems crucial to correct for the participants height if you’re going to measure the genetic influence on the volume.

    Otherwise you might end up with an effect on height but no “real” effect on the intracranial volume. Especially if this gene is already known to account for height differences.

  5. Ed

    Is there a strong correlation between head size ( hat size etc.) and intracranial volume?

  6. Dominic L

    Meta-analysis pulls out the interesting finding for hippocampal volume. But I don’t know if the IQ thing will pan out. What was the P value on that SNP associated with IQ?

    Also, with Bastian, hope that the authors had the good sense to adjust volumes for height. Given that it was a meta-analysis, some may have and others may not.

    If I had to bet, I do think the intracranial volume thing will replicate, but the IQ part probably will not.

    Has anyone tried tackling IQ using genetic risk scores or other prediction models like the folks who publish on height do? A better study might be measuring native intelligence in small children. Not an easy study to do: DNA and IQ data from the under 4 set.

  7. Miley Cyrax

    Phew. Good thing it’s African populations that have the highest frequencies of allele C, else the media and public would immediately decry this kind of research.


    Ed, this kind of stuff is easily findable via a quick Google search. It’s been found quite a few times to be a .3 to .4 correlation.

  8. re: IQ. after 24 hours i would go “short” on this being a true positive. got some information which suggests this is probably not going to replicate.

    The biggest concern in my view in this study would be that a merely ancestry informative marker with neutral effect on intelligence is picking up of intelligence differences attributable to histories of immigration

    well, as i said, the authors are not stupid :=) the samples were european.

  9. Ed

    @Miley Cyrax

    Lmao. I know I did a google search like 30 minutes after I posted that. Found this:

    “Even though head size also depends on factors such as the muscularity of the head and thickness of the bone, it’s very likely that a bigger head means a bigger brain.”

    Would you mind posting your source for the correlation of .3 or .4 between “head size” and “brain size”?
    For my own personal reference.

  10. toto

    The biggest concern in my view in this study would be that a merely ancestry informative marker with neutral effect on intelligence is picking up of intelligence differences attributable to histories of immigration

    Population structure is a biatch, but… surely someone has come up with a way to control for that? Even something as dumb as partial correlation between gene of interest and trait, controlling for the (PCA’ed) rest of the genome might provide some info… Though of course you might need humongous Ns to get any signal.

  11. I don’t recall what the correlation between head size and brain size as measured by MRI is, but the correlation with g is ~0.2, and ~0.4, respectively. See

  12. Martin

    Is it 9 mm3 or as stated in the New Scientist 9 cm3? I don’t have access to the original publication….

  13. sofiarune

    Have you looked into Rushton’s background at all? He has undermined pretty much any trust that can be put into his research. Gotta love an “academic” who thinks the French Army Surgeon and Penthouse Forum are legitimate sources for data.

    The number 9,006.7mm3 in this article is correct (it’s a direct cut and paste from the original article which I also have). 9,006.7 mm3 which is about ~9 cm3.

  14. DKshadow

    @sofiarune Taking your mischaracterizations of Rushton beyond YouTube I see. The paper Emil cited wasn’t just authored by Rushton, Davison Ankney is a coauthor. It’s not JUST Rushton’s research. So you’re attacking the credibility of Davison Ankney too.

    Also the brain size/GMA correlation in Rushton’s paper is fairly consistent with Michael A. McDaniel’s meta-analysis:


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar