Saturday, March 2, 2013

It's not because research is cheaper there


Robert Plomin on the genetics of various mental traits (source)


A Chinese research team is looking for genes that explain why IQ is higher in some people and lower in others:

Studies show that at least half of the variation in intelligence quotient, or IQ, is inherited. But while scientists have identified some genes that can significantly lower IQ—in people afflicted with mental retardation, for example—truly important genes that affect normal IQ variation have yet to be pinned down.

The Hong Kong researchers hope to crack the problem by comparing the genomes of super-high-IQ individuals with the genomes of people drawn from the general population. By studying the variation in the two groups, they hope to isolate some of the hereditary factors behind IQ. (Naik, 2013)

The head of the team, Zhao Bowen, believes this question has not been resolved because it is too controversial. “People have chosen to ignore the genetics of intelligence for a long time," said Mr. Zhao, who hopes to publish his team's initial findings this summer. "People believe it's a controversial topic, especially in the West. That's not the case in China" (Naik, 2013).

Perhaps. But there is another reason: the apparently large number of genes involved and the relatively small effects of each one. This was the conclusion of Robert Plomin, a behavioral geneticist at the Institute of Psychiatry in London:

Failing to find genes for intelligence has, in itself, been very instructive for Plomin. Twin studies continue to persuade him that the genes exist. “There is ultimately DNA variation responsible for it,” he says. But each of the variations detected so far only makes a tiny contribution to differences in intelligence. “I think nobody thought that the biggest effects would account for less than 1 percent,” Plomin points out. 

That means that there must be hundreds—perhaps thousands—of genes that together produce the full range of gene-based variation in intelligence. (Zimmer, 2008)

This should be no surprise. Natural selection doesn’t act on genes, at least not directly. It acts on phenotypes—the flesh-and-blood outcomes of genes. Selection for intelligence will thus affect any gene that has some kind of intelligence-boosting effect.  This point has been made by Linda Gottfredson, a psychology professor at the University of Delaware:

[...] within-group ('individual") differences in intelligence will involve 1000s of genes of small effect, so we can expect that for between-group differences too. Many of the genes will not be specific to intelligence per se but influence broad physiological processes that affect brain structure and function. This would include cardiovascular fitness and much more. (Go exercise, guys!)

I read that perhaps half our genes are expressed in the brain. If half of our segregating genes are too (the 0.1% on which humans are estimated to differ), that's still 1.5 million base pairs or "SNPs" (of 3 billion total).

This indicates the challenge, even if we ignore other important genomic differences (e.g., number of times a given segment of the chromosome is repeated, like a stutter).

[...] This is not to say, of course, that we can't pin down heritabilities for various mean group differences (we could right now if researchers were willing) or that we won't be able to identify numbers or classes of genes on which groups differ most. But it's looking unlikely that we'll be able to pinpoint a list of specific genes that explain much of the normal variation in g, either within or between groups. (Gottfredson, 2013)

Interestingly, Robert Plomin is mentioned as one of the people involved in the Chinese project. Has this research been offshored to a country where the intellectual climate is less restrictive? In addition, since Plomin is aware of the limitations of this kind of study, he might know something that the rest of us don’t. Perhaps among the many genes with small effects there are a few with big effects …

References

Gottfredon, L. (2013). H-bd discussion list, February 22, 2013

Naik, G. (2013). A genetic code for genius, The Wall Street Journal, February 15
http://online.wsj.com/article/SB10001424127887324162304578303992108696034.html?mod=WSJ_hp_mostpop_read

Zimmer, C. (2008). The search for intelligence. Scientific American, October, pp. 68-75.

 

29 comments:

Ben10 said...

In classical genetic, the endeavor would be indeed meaningless, since the finding of thousands of genes each involved in very small amount would be practically impossible.

But computational approaches for analyzing high-throughput, multi-dimensional genomic and epigenetic data can do the job, and from there, computer models of interacting network of genes and epigenetic environmental factors can be build. From there, it's not impossible to implement genetic and epigenetic measures that could lead to a global IQ increase in the population.
After all, chinese policies already include 1 child (boy) policy, why could they not also include 'variable child number policies' according to the parents likelyhood of delivering high IQ children?
But i don't believe in that because high IQ population is uncompatible with authoritarian regimes. Chineses are plenty smart already, why would their regime want them smarter, more individualistic, in short becoming potential dissidents, more likely to challenge the rules?
Unless they cam also find a set of conditions for submissivity.

And that's the same in the West by the way. More and more people are cosncient that today's democracies are a farce and voting a complete waste of time. Why would the occidental oligarchic regimes want more people aware of that?

Unless the elites, in China and the West, want to keep the secrets for higher IQ for themselves and their castes. That's Huxley's description in Brave New World, with its Alphas and their 'Betas' aka 'Slaves' in common language.

Anonymous said...

But there is another reason: the apparently large number of genes involved and the relatively small effects of each one. This was the conclusion of Robert Plomin, a behavioral geneticist at the Institute of Psychiatry in London

This would be consistent with Gerald Crabtree's recent work. At least as far as the view that many genes are involved goes:

http://www.telegraph.co.uk/science/science-news/9672790/Civilisation-is-making-humanity-less-intelligent-study-claims.html

"Prof Gerald Crabtree, a developmental biologist at Stanford University, explained in the Trends in Genetics journal that a mutation in any one of 2,000 to 5,000 particular genes could lower our intellectual and emotional ability."

Sean said...

I doubt mutations in any of 2-5000 genes could noticeably lower IQ. Gregory Cochran thinks mutations from older fathers in polygynous populations have lowered IQ, but it seems to me that if individual genes with "the biggest effects would account for less than 1 percent" (as Plomin says); then most mutations in most genes get buffered out of expression, and genes with a big effect are few in number.

Anonymous said...

it seems to me that if individual genes with "the biggest effects would account for less than 1 percent" (as Plomin says); then most mutations in most genes get buffered out of expression, and genes with a big effect are few in number.

Isn't Plomin saying that the genes with the biggest effects have effects of less than 1%, not that less than 1% of the genes have the biggest effects?

Anonymous said...

"This is not to say, of course, that we can't pin down heritabilities for various mean group differences (we could right now if researchers were willing) or that we won't be able to identify numbers or classes of genes on which groups differ most."

How would researchers go about doing this? Is there a way you (Peter) or someone can give a layman a succinct explanation?

gcochran said...


In considering the variance explained by a given allele, you must consider frequency as well as its effects. A rare variant (or a fixed variant) will not account for much of the variance even if it has a large effect. Variants have their maximum effect on variance when their frequency is 50%.

Next,if you find that no variant explains more than a small amount of the variance in a given trait in a particular population, you don't know that this will be the case in a different population.
The genetic architecture of height looks to be quite different in Pygmies than in Europeans.

Sean, we already know of many single-gene mutations that materially decrease IQ. You should know!

The Protestant Elite said...


Sean, we already know of many single-gene mutations that materially decrease IQ. You should know!


As a holder of many of those mutations, your insult probably went over his head.

Sean said...

gcochran, I believe Pygmies are unusually short because they have been selected for that. Differences in the genetic architecture of intelligence between populations would be surprising, unless one population had been under directional selection for lower IQ, and the other for higher IQ

I took Crabtree to be claiming that we all have at least 2000 genes in which any mutation would produce a noticeable attenuation of mental performance. Unless the "many" in "many single-gene mutations that materially decrease IQ" means 'many thousands', what I take to be Crabtree's assertion that mutation at any of 2-5000 genes will have a noticeable effect on mental qualities remains dubious, by my way of thinking.

Sean said...

Kevin Mitchell takes Crabtree to the woodshed; here.

'Mutations causing intellectual disability dramatically impair fitness, explaining why so many cases are caused by de novo mutations – the effects are often too severe for them to be inherited [5]. Furthermore, this selective pressure extends into the normal range of intelligence, as described by Deary: “One standard deviation advantage in intelligence was associated with 24% lower risk of death over a follow-up range of 17 to 69 years… The range of causes of death with which intelligence is significantly associated… include deaths from cardiovascular disease, suicide, homicide, and accidents”'

Sean said...

Could background selection make it difficult to identify genes affecting g.

n/a said...

"In addition, since Plomin is aware of the limitations of this kind of study, he might know something that the rest of us don’t. Perhaps among the many genes with small effects there are a few with big effects"

At least within Northern Europeans, no surprise is likely. The study has been done. You can estimate breeding values for IQ from microarray data right now. The estimates just won't be very accurate until sample sizes increase.


"unless one population had been under directional selection for lower IQ, and the other for higher IQ"

That could be the case.

Juoni said...

I believe Pygmies are unusually short because they have been selected for that.

Probably not:

"Short stature in African pygmies is not explained by sexual selection. Evolution and Human Behavior Vol. 33, Issue 6, Pages 615-622"

Peter Fros_ said...

Ben,

The same question is on my mind. Can this research team identify so many genes with such small effects and puzzle out the interactions among all those genes?

It's doubtful whether China's one-child policy is having any impact nowadays. Fertility rates are actually lower in Taiwan and Singapore, where there is no such policy.

You don't need an advanced understanding of genetics to select for intelligence. The Chinese state could simply create a special family allowance for young couples who do well on their university admission exams.

Anon,

Yes, and this seems to be the case for almost any complex trait, e.g., height.

Sean,

There is a measurable paternal age effect, but it doesn't seem to be a factor in population differences, at least not a major factor. Paternal age has been significantly higher in Western Europe than elsewhere for about a millennium, yet mean IQ is, if anything, higher there. Maybe I'm self-conscious. (In my family tree, there has been a longrunning tendency towards late marriage).

Anon,

Yes, if we look at differences in IQ from one person to the next, no one gene contributes to more than 1% of the variability.

Greg,

I overlooked that point. Yes, the bigger the effect of an allele, the less frequent that allele will be. There might be a lot of alleles with big effects, but they probably occur most often in semi-autistic individuals.

Some populations, like Ashkenazi Jews and Parsees, may have more of these big-effect alleles because selection for intelligence has been so strong and so recent (relatively speaking). The Tay-Sach's allele is a case in point. It seems to boost intelligence, but at the cost of causing early death in homozygotes.

Protestant,

Uh, it went over my head too.

Sean,

I think Greg was referring to the time frame of selection. If there has been strong selection over a short time, you'll get more big-effect alleles (and also more undesirable side-effects).

n/a,

I know. Maybe this isn't a serious methodological barrier. We'll soon know (the results will come out this summer).

Juoni,

What about natural selection?

Sean said...

Peter, thanks for the clear explanation of GC points, and the fascinating suggestion about high value alleles.

Re mutations: some are adaptive, for the fathers at least. Pedigree dog breeders who get a champ often fix it as a 'type' with inbreeding, thereby increasing the effect of mutations (that aren't too serious). A problem with pushing too far the dogs move oddly, even through they look good. (Compare the effortless gait of a wolf). Black Africans move well. I wonder about buying of wives; perhaps before there were cattle or other means of exchange, African fathers were younger.

"You don't need an advanced understanding of genetics to select for intelligence."

Life selects. On average, those who do not believe in a genetic basis for individual or group differences in intelligence, are significantly more intelligent; brainpower's function is not discovering truths, it's to survive and reproduce by getting on in life. That's why western society is currently run by the people it is, and will be run by their children in the future.

Maybe those high IQ individuals who believe in individual and group hereditary differences are driven to make their socially dysfunctional conclusions because their high intelligence has a somewhat unusual architecture; it tends to be conferred by high value alleles that also confer semi-autistic traits.

So, US elites might be wise not to want their children to get any high value alleles that studies of maths prodigies might through up, as these are such a double edged sword.

Anonymous said...

Life selects. On average, those who do not believe in a genetic basis for individual or group differences in intelligence, are significantly more intelligent; brainpower's function is not discovering truths, it's to survive and reproduce by getting on in life. That's why western society is currently run by the people it is, and will be run by their children in the future.

They’re lying to themselves, which is the ultimate fate of those who rely on con artistry too much. They’ll be toast in any real war because they won’t be able to sort out their lies from nature's truths.

Anonymous said...

BTW, speaking of Parsees and the Ashkenazim, I assume that in populations that have much been naturally (and not just sexually) selected for intelligence, you would expect a lower correlation between g and other fitness indicators. Has anybody considered this?

Anonymous said...

"Perhaps among the many genes with small effects there are a few with big effects …"

yep

Saying there's only 1000s of genes with tiny effects would be a neat way of stopping people looking for those few others.

This may be overly cynical but after Gould, Lewontin etc i think being overly cynical is perfectly reasonable.

Sean said...

It is going to be difficult to identify genes using DNA samples from high-IQ individuals with high STAT scores, physics doctorates ect, because collectively those people just have g of a very high order, and the rare high value genes would tend to be different in them.

But Niak says they're analysing DNA from mathematically gifted people only. Now a few specific high value genes are likely to be particularly useful for advanced maths; maths prodigies are more likely to share specific high value genes, so identification of one or two rare alleles is quite likely.

Anonymous said...

Where'd that Gottfredson quote come from ?

Anonymous said...

"intelligence (difference) between populations would be surprising" -- NO, the exact opposite is true: what would be surprising is exactly equal population means for all groups -- it would require that every time a bright eskimo died, a dull pygmy is born, in order to keep the population means exactly equal. to expect mu = mu = mu = mu, etc. is ludicrous!

Sean said...

'mu' is 'the population mean or expected value in probability and statistics' according to wikipedia.

Oh dear! panjoomby is afflicted with the mechanistic cognitive style so common in engineers, mathematicians, physicists, the children of Silicon Valley, (and by his own account), Bill Hamilton:- "a person who ... believes he understands the human species in many ways better than anyone and yet who manifestly doesn't understand in any practical way how the human world works - neither how he himself fits in and nor, it seems, the conventions. [...] unaffected by all the propaganda they have failed to hear, [they] see further into the true shapes that underlie social phenomena."

But it doesn't matter what painfully perceptive hard scientists say about the differences in mu and alleles, because brilliant biologists at elite universities deploy rationalism (arguing from first principles) to prove that conventional beliefs must be correct, and reasoning or results suggesting otherwise must be wrong. "HE says he thinks that no significant genetic differences will be found between races because of his belief in the efficiency of natural selection. Just as selection turns out to have pruned away most disease-causing variants, it has also maximized human cognitive capacities because these are so critical to survival. “My best guess is that human intelligence was always a helpful thing in most places and times and we have all been under strong selection to be as bright as we can be,” he said.

This is more than just a guess, however. As part of a project on schizophrenia, Dr. Goldstein has done a genomewide association study on 2,000 volunteers of all races who were put through cognitive tests. “We have looked at the effect of common variation on cognition, and there is nothing,” Dr. Goldstein said, meaning that he can find no common genetic variants that affect intelligence. His view is that intelligence was developed early in human evolutionary history and was then standardized."


And to anticipate what might be thought about a special allegiance on the part of certain biologists; that attitude is as American as apple pie, or American transcendentalism's repugnance to Darwinism, which goes way back (see here).

Anonymous said...

rejecting "mu = mu" (null hypothesis) may mean i worship science, & by worshipping science i may be blind to the idea that some null hypotheses are silly... not to argue from authority (!) but i'm reminded of jon haidt's research on morality - how it "binds & blinds" so that holders of some beliefs treat the beliefs as sacred (e.g., there can be no sex differences, b/c that would be non-egalitarian) & therefore create grandeloquent non-occam's razor type elaborate environmental theories b/c certain blunter hypotheses are simply not to be considered. we all do this - e.g., the right is blind to certain things - say, climate change - the left may hold egalitarianism so sacred that they must ipso facto believe when a bright eskimo dies, a dull pygmy is immediately born (my shorthand for mu = mu:) a problem for human science is if most profs lean a certain direction, then certain hypotheses are given short-shrift or dismissed outright (so we get IQ-deniers, measurement deniers, mean group difference deniers, the incorrect cry of "but those tests are biased" or variability deniers). since we pre-lean in a certain direction, we're prone to find the one study that says what we believe (& be satisfied with it!) & not regard the studies that do not. haidt's findings (which we may even agree on): intuition comes first, & then strategic reasoning is simply our post-hoc way of defending our original intuition.

Sean said...

I am not aware of any modern elected politician even suggesting that there may be group differences though they'll say just about anything. Politicians operate in a mentalistic way: they understand their job is not to educate society about the need to take hard science as a guide to human potential. No, it's to affirm ideals and values as the determining factors that shape society, which sounds noble and heroic .

Anonymous said...

Hasn’t ANYBODY here read Millers’ brilliant “Mating Mind”, where he argues that human intelligence beyond a certain level has much been increased due to sexual, rather than natural selection, and thus one would exactly EXPECT wide differences between people (unlike with natural selection, where there is little variance in what is universally better), as well as that g correlates with other fitness indicators (it also answers the question of many intelligent people, “what is intelligence good for?”). And (methinks), one should consider this in light of the fact that other differences between populations are probably due to sexual selection, and the fact (just thinking out loud) that the most intelligent populations tend to have hotter females than males. Anyways, as I said above, one would expect that in populations that have been under natural selection for high intelligence, there should be a lower correlation between g and other fitness indicators. Anybody aware if this has been considered?

Anonymous said...

good call - the mind beyond survival needs - is a peacock's tail! that allows for more accurate assortative mating/homogamy(smarties can be more assured they're selecting smarties:) & i hadn't ever considered your idea before of smarter populations having hotter females - you're spot on - that's fascinating! i'm sure there's at least a modest correlation between female hotness & intelligence (b/c extremely low IQ females would rarely be hot) so the peacock's tail strikes again!

Peter Fros_ said...

Sean,

Except that our elites had a very different attitude not so long ago. It was really the end of the Cold War and the increased shift to globalization that changed everything. We're entering a period of "gangster capitalism" not unlike that of Russia during the Yeltsin years.

Anon,

I'm not sure I follow your argument. Why would the correlation be lower?

Anon,

I suspect there are some genes with big effects. I also suspect that they have other big effects that are less desirable (autism, excessive shyness, etc.)

Anon,

The quote is from the h-bd list. It's a group dedicated to discussion of human biodiversity.

Panjoomby,

It would be surprising if all human populations had exactly the same mean IQ. Forget differences in selection pressure, even founder effects would produce some difference.

Anon,

"unlike with natural selection, where there is little variance in what is universally better"

And why is high intelligence universally better? Some societies have little use for abstract reasoning. The emphasis is on learning certain rules of life without pondering the "why" of it all.

Anonymous said...

"Just as selection turns out to have pruned away most disease-causing variants, it has also maximized human cognitive capacities because these are so critical to survival."

This only makes sense if the energy costs of a low-functioning and high-functioning brain are the same. If they're not then it makes sense for evolution to aim at the minimum level of average intelligence to survive in a particualr environment. (Where the average provides enough outliers for the witchdoctors and chiefs.)

If so then you'd expect the large-effect genes to mainly pop up when they're being specifically selected for.

.
"I suspect there are some genes with big effects. I also suspect that they have other big effects that are less desirable (autism, excessive shyness, etc.)"

Yes, me too hence why i'm more interested in the genetic load reduction idea.

M said...

This only makes sense if the energy costs of a low-functioning and high-functioning brain are the same. If they're not then it makes sense for evolution to aim at the minimum level of average intelligence to survive in a particualr environment. (Where the average provides enough outliers for the witchdoctors and chiefs.)

If so then you'd expect the large-effect genes to mainly pop up when they're being specifically selected for.


With cognitive capabilities (which may not be intelligence rather than any kind of intensive process) - its correct evolution should tend to work to maximize the efficiency of the brain. Efficiency means power relative to cost - the most power relative to lowest cost should maximize.

Maximizing cognitive capabilities in ways which involve increases resource consumption - which may be in the form of energy use or size, both of which cost resources - that will not happen unless the selective pressures are correct.

So evolution will and will not work to maximize cognitive capabilities.

Also, one thing to note is that, as Cochran has noted before, there are higher returns to cognitive efficiency with higher cognitive resource consumption, so larger brains and brains which habitually use more energy (burn more glucose) will both tend to behave more efficiently (use smaller networks and less glucose) for the same task (relative to smaller and colder running brains). Which is exactly what we find. The co-evolution of efficiency and resource intensiveness would lead to efficiency promoters working better in larger and hotter running brains.

Anonymous said...

The video...

1) ulcers from stress? huh, doesn't he know it's h. pylori, not stress, even though, yes, there's a genetic component probably?

2) monozygotic twin concordance for male homosexuality is not approx 40%, as he says (he refers to the work of Mike Bailey), but approx. 20% according to the largest and latest twin study that Baily now references (I think the sample was from the New Zealand twin registry)

Is this video old?