Very interesting! One thing I can't tell from all the summaries and writeups, though, is how the carriers of rs174575 stack up to the ancestral allele in the absence of breastfeeding. In other words: are we looking at a situation where two populations have the same baseline IQ, but the variant carriers end up with a seven-point edge if breastfed? Or are we looking at a situation where the carriers need to be breastfed in order to attain the same level of IQ as the baseline population? Or something in between? Based on other information I'd assume the first case, but I can't tell for sure.

Bbartlog:

You're correct in that it is the former case.

In the study itself, there were two cohorts: a British cohort and a New Zealand cohort. In both cohorts, presence of the C allele (as either CC or CG) was associated with a hike in IQ by 6.4 and 7.0 points (from around 99-100 to around 106) in the two samples. But those without the C allele (GG) had mean IQs of 99.5 in Britain and 100.3 in New Zealand.

Incidentally, I plugged rs174575 into the ol' Hapmap, and it seems that the C allele has a frequency rate of 72% among Utah whites, 73% among Yorubans, 89% among the Chinese, and 90% among the Japanese.

Assuming these numbers are representative, universal breastfeeding alone would lead to a gap of 1.12 to 1.225 in IQs between East Asians and whites, favoring East Asians.

On the other hand, since whites and blacks have nearly the same C allele frequency (assuming, again that these numbers are representative of white and black populations), but whites breastfeed their babies at higher rates than blacks (60% vs. 30%, according to Steve), this is pretty solid evidence that increasing the percentage of black women who breastfeed their children will have a diminishing effect on the black white IQ gap.

Since I'm already blowing off work (damn you, Gene Expression!) I might as well do the calculations. Let's see... The 30% difference in breastfeeding is responsible for between 1.4 and 1.53 points in the black-white IQ gap! (I got these figures by multiplying the frequency rate of the C allele in Yorubans with the percentage difference in breastfeeding between blacks and whites with the effects of breastfeeding on the IQs of C allele carriers.)

Steve Sailer is vindicated!

Ha Mark, I beat you to the punch at Future Pundit!

http://www.futurepundit.com/arch...ves/ 004759.html

Who can read now, huh? :)

I can.

Rob,

Yes, you did indeed beat me to the punch. Very impressive. Or rather, it WOULD BE impressive, except that your assumption that:

Children with the C SNP gain 7 points over both people with the G/G genotype and people with C/x genotype who are not breastfed, right?

IS WRONG. Children with the CC genotype AND the CG genotype gain 7 points over people with the GG genotype only. (I.e. you only need one C allele to get the seven points). That may be why our results differ slightly. (I'm too tired at this late hour to see if that is the case.)

I knew I didn't spend ten bucks downloading that goddamn article for nothing. :-)

Oh, one could misintepret that: I certainly wasn't clear enough. People with a C gain on people who are GG and people who have a C AND are not breastfed. In the math I used 7 points for both CC and CG.

Our results may differ because you have the 6.4 point boost for heterozygotes, whereas I used 7 points for both CC and CG. You spent $10 for .6 points :)

Also, in my laziness and optimism, I upped the breast-feeding rate for both whites and blacks to 100%, you equalized them both to 60%. And that is probably more achievable.

Thanks Jason, I would also like to add that this means two wonderful things.

1) Part of the black-white IQ gap is, depending on exactly how one defines it, either environmental or amenable to environmental intervention.

2) IQ for both whites and blacks can go up at least a bit even in developed, first world countries.

I'm sure that you are not fond of Robert Lindsay at this point, but he mentioned, and I knew before, that IQ correlates to several physiological traits, like frontal lobe volume. To the extent that we can tell, have these things been going up with the Flynn effect?

And a possibility of bad news: I didn't spend $10 on the study. At what ages were IQ measured? Is it possible that the gains decline or disappear in adulthood?

G Prize Represent!

fyi, I (and I assume other people who post here) would be more than willing to send anyone a copy of any paper we post about. just shoot an email.

Yeah, added to gnxp forum files: 'breastfeedIQ.pdf'.

Rats. Well, next time I'll know better than to blow ten bucks in a rush of geek euphoria.

Anyway, to answer Rob's question, for the New Zealand sample, children were tested at ages 7, 9, 11, and 13. For the British kids, IQ was tested at age 5, according to the study.

Anyone want to weigh in on the selection signature for this SNP? I've looked at haplotter, but not being all that familiar with the various calculated stats my naive conclusion is that it looks sort of selected for, but not enough to pop out on in any sort of statistical filtering effort. More generally, if a non-ancestral allele reaches 70-90% frequency in the population and is shown to have a significant function (not just neutral genetic dross), can we confidently say it's been under positive selection? I guess we still can't rule out the possibility that it's the result of selection on some nearby chunk of genetic code...