DTNBP1 gene and racial IQ differences

Based on the study linking various SNPs (single nucleotide polymorphisms) of the DTNBP1 gene, I created the following chart summarizing data from the DTNBP1 study and the HapMap Project. (Sorry about the horizontal scrollbar.):

SNP	Common allele	Uncommon allele	IQ benefit	Asian freq	White freq	Nigerian freq	Asian IQ	Nigerian IQ
rs2619539	G	C	3.5	85.39%	69.49%	51.67%	0.5565	-0.6237
rs3213207	A	G	-2.86	1.12%	20%	5%	0.539968	0.429
rs1011313	G	A	-0.21	46.07%	13.33%	0%	-0.068754	0.027993
rs2619528	G	A	-6.41	15.12%	34.48%	63.33%	1.240976	-1.849285
rs760761	C	T	-8.02	14.44%	32.20%	63.33%	1.424352	-2.496626
rs2619522	T	G	-7.51	15.56%	33%	62.71%	1.309744	-2.231221
rs2619538	A	T	6.52	3.41%	81.36%	88.14%	-5.08234	0.442056
							-0.079554	-6.301783

The chart above shows the IQ benefit of carrying at least one uncommon allele, and the percent of population having at least one uncommon allele. The righthand columns calculate the IQ difference with respect to the white population.

The net result shows -6.3 IQ points for the Nigerian population group compared to the whites, and -0.1 IQ points for Asians compared to white. Asians are supposed to have higher IQs than whites, so I suspect that other intelligence genes such as the CHRM2 gene, will be found to boost Asian IQ.

The really strange SNP is rs2619538, because that shows -5.1 IQ for Asians compared to whites. Most of the other SNPs show a higher IQ for Asians and a lower IQ for Nigerians. I wonder what else that SNP does?

DTNBP1 explains 1/3 of the black-white IQ gap.

Looking at the allele frequencies, all but one of the SNPs have the white population group in the middle compared to the Asian and Nigerian population groups, tending to confirm J. Philippe Rushton's r/K selection theory.

Human biodiversity deniers say "races have 99.9% same genes blah blah blah more variance withing races than between races blah blah blah," but as we see, it only takes a few SNPs out of billions to to explain 1/3 of the black-white IQ gap.

I will be studying some of the other IQ genes in the future to get a better picture. I suspect that half of the Asian-white-black IQ gaps will be explained by currently known IQ genes.

UPDATE 1/30/2010

More than two years after I wrote this, this post suddenly appeared on the front page of reddit.

I was just playing around with the HapMap project when I wrote this post. Although I do believe that the ability to accurately predict a person's IQ based on their genotype will be possible in the not-too-distant future, there aren't yet enough genes identified in the literature for an armchair geneticist to determine the average racial IQs differences between whites, Asians and Nigerians.

HOWEVER, what the existing research and the HapMap project demonstrate is that (1) various SNPs have been identified which are associated with brain function and intelligence; and (2) the frequency of occurrence of these SNPs differ dramatically between races. I am sure that there will soon be more compelling DNA-based proof that the average black isn't as intelligent as the average white. The non-DNA-based proof is compelling.

You should read my post Response to comments on race and intelligence in which I probably responded to the comment you wanted to write.

UPDATE 2/1/2010

I already admitted I was just over-exuberantly playing around with the HapMap project and I’m now aware that the method was flawed. At the blog Gene Expression, you can read a more detailed explanation of why this specific analysis doesn’t prove much.

But once again, the bigger point of it all is that (1) there are SNPs associated with brain function; and (2) these genes have a very different distribution in different racial gene pools. This demonstrates that there’s a sound basis for the theory that IQ differences and other behavioral differences between races has a genetic explanation.

Comments

[Jason Malloy]

One idea is to sidebar a continually updated post.

Just throw CHRM2 up on the same chart, and keep adding numbers as you find genes.

Here are a couple references to get you started.

You'll begin to find a lot of complexity: epistasis, mediation by specific environmental effects (as purported in the breastfeeding study), or stages of development (genes which boost IQ in seniors but not in children). And of course most of these studies just aren't replicated.

But a post that can be used as an evolving scratchpad might be useful to organize information as it comes along.

[Jason Malloy]

Sorry for the double post. Here's a brand new study I just found:

"Recently, a family-based genetic association was reported between variation in intelligence quotient (IQ) phenotypes and two intronic variants on the SNAP-25 gene... Two new variants in intron 1 (rs363043 and rs353016), close to the two previous reported variants (rs363039 and rs363050) showed association with variation in IQ phenotypes across both cohorts."

It doesn't say how much IQ in the abstract. I'll try and get this study and upload it to gnxpforum by tomorrow night.

[Gil]

8(

[SFG]

Isn't there a correlation problem in that this gene may be linked with others having the same effect? IE, maybe it doesn't explain the whole 6 point difference?

[SFG]

I'd be surprised if half the gap were explained by currently known genes. Honestly, we really know very little.

[Robert Hume]

Who are the "Asians"? As I remember it is NE Asians, e.g. Chinese, Japanese, who have higher IQs. Others, e.g. Malays, Filipinos, Vietnamese, Cambodians, have lower IQs.

[Half Sigma]

SFG: "I'd be surprised if half the gap were explained by currently known genes. Honestly, we really know very little."

Some researchers have been intentionally searching for IQ related genes, so the ones that are found are the ones that have the largest impact on IQ (because they are the ones that are easiest to find). Based on the 80/20 rule, if there are 100 SNPs affecting intelligence differences between people, 20 of those SNPs probably account for 80% of the differences.

[Half Sigma]

Who are the "Asians"?

50% Chinese and 50% Japanese.

The whites are people of European ancestry living in Utah.

[Marc]

If you do as Malloy suggests, which I think would be a great idea, then you should take care to note which alleles have been associated with increased IQ based on more than one study. At least one allele that was originally associated with IQ (I forget which one, but it's not the one you reference above) in the mid 90's didn't hold up. So you could have a list of "confirmed" alleles and "tentative" alleles, and a "confirmed" effect and a "tentative" effect.

[Rob]

And HS, don't forget that, eg CHRM2 affects performance IQ, and to get the full-scale IQ impact, you'll have to divide the performance IQ by 2.

[Fly]

"Some researchers have been intentionally searching for IQ related genes, so the ones that are found are the ones that have the largest impact on IQ (because they are the ones that are easiest to find). Based on the 80/20 rule, if there are 100 SNPs affecting intelligence differences between people, 20 of those SNPs probably account for 80% of the differences."

SNP's that significantly increased IQ should sweep a population quickly. Even rare matings between geographically separated populations would pass on the allele. So alleles of large effect should either be very recent or balanced by significant negative effects. Due to selection there should be fewer alleles of large effect than predicted from a power law.

In my opinion the SNP differences found so far don't begin to account for the IQ ranges seen in even one racial group. I suspect non-additive genetic effects and structural genome differences are important.

[Jason Malloy]

Ok folks, I've added the SNAP-25 and CHRM2 PDFs @ gnxpforum. But it won't let you access them without first being registered and logged in.

Thanks be to Agnostic for SNAP-25.

[Fog]

Before getting too excited about this, please make sure that the impact of those genes on IQ is solidly established AND tell us how much uncertainty (standard deviation) there is with the point estimates. They are probably based on small samples. This is the bit here I'm most suspicious with.

[Henry Harpending]

Nice summary. It would also be useful and (maybe?) easy to post which allele is the ancestral one at each SNP.

[Jason Malloy]

The information for Dr. Harpending's suggestion can be acquired by an SNP search at PubMed.

For instance the ancestral allele for rs2619539 is G.

[SFG]

"Based on the 80/20 rule, if there are 100 SNPs affecting intelligence differences between people, 20 of those SNPs probably account for 80% of the differences."

The 80/20 rule applies to human relations such as business and sociology. You sure it applies to genes and proteins?

[Half Sigma]

After thinking more about it and reading the excellent comments, perhaps the 80/20 rule doesn't apply to the IQ genes causing racial differences in intelligence.

[Rob]

What's the mathematical origin of the 80/20 rule? For a normally distributed random variable, is 80% of the area under the curve under the right hand 20% of the variable?

Like Fly says, alleles with totally positive results will sweep. The only ones we'll find with big population differences will be ones that have big drawbacks, at least on occasion. OTOH, those are the ones most likley to be identified by medical research.

"DTNBP1 explains 1/3 of the black-white IQ gap."

No It does not. The B-W gap in the US is 14 points, but the Nigerian White gap is ~30 points. If I'm not too lazy, when I get back from the gym, I'll work out the probable distribution and gap in the US African-American population.

Are the IQ-boosting versions dominant or additive?

[Fly]

"After thinking more about it and reading the excellent comments, perhaps the 80/20 rule doesn't apply to the IQ genes causing racial differences in intelligence."

As for my comment...I wasn't really disagreeing with your suggestion. I was only trying to improve the model.

As an analogy: A normal curve is a good first approximation to the US population IQ distribution. However, the far left side needs a bump to account for severe birth defects and the far right tail is fat (probably because of assortative mating and because it is the sum of distributions for different genders and races).

I believe power laws are common in genetics/evolution/biology. There has been research concerning bacterial adaptation that suggests mutations and allele variation follow power laws. I thought your intuition about an 80/20 rule was good. I was only pointing out that the left side of the 80/20 curve might be "flattened" due to selection.

[Rob]

No It does not. The B-W gap in the US is 14 points

Oops, typo, meant 15 points.

[ben g]

"DTNBP1 explains 1/3 of the black-white IQ gap."

BS. only if it was the only gene that existed.

[Gaham Adams]

Add the following: CHRM2 gene
SNAP-25 gene
DTNBP1 gene / DYSBINDIN-1 gene
APP gene
APOE gene
FADS2 gene
PCNT gene
IL1RAPL1 gene
ASPM gene
DARPP-32 gene
and
CAG repeats