Genome-wide association studies establish that human intelligence is highly heritable and polygenic
General intelligence is an important human quantitative trait that accounts for much of the variation in diverse cognitive abilities. Individual differences in intelligence are strongly associated with many important life outcomes, including educational and occupational attainments, income, health and lifespan. Data from twin and family studies are consistent with a high heritability of intelligence, but this inference has been controversial. We conducted a genome-wide analysis of 3511 unrelated adults with data on 549 692 single nucleotide polymorphisms (SNPs) and detailed phenotypes on cognitive traits. We estimate that 40% of the variation in crystallized-type intelligence and 51% of the variation in fluid-type intelligence between individuals is accounted for by linkage disequilibrium between genotyped common SNP markers and unknown causal variants. These estimates provide lower bounds for the narrow-sense heritability of the traits. We partitioned genetic variation on individual chromosomes and found that, on average, longer chromosomes explain more variation. Finally, using just SNP data we predicted ~1% of the variance of crystallized and fluid cognitive phenotypes in an independent sample (P=0.009 and 0.028, respectively). Our results unequivocally confirm that a substantial proportion of individual differences in human intelligence is due to genetic variation, and are consistent with many genes of small effects underlying the additive genetic influences on intelligence.
Pessimism of the Intellect, Optimism of the Will Favorite posts | Manifold podcast | Twitter: @hsu_steve
Tuesday, August 09, 2011
Intelligence: heritable and polygenic
Below are the title and abstract of the paper I hinted at in this earlier post. Although the study failed to find any specific loci that are associated with intelligence, a global fit showed that a significant chunk of the heritability expected from twin and adoption studies is accounted for by SNPs. In other words, genetic similarity is correlated with similarity in g score, even though we don't know which genes are specifically responsible. The results of the study were expected from what we already knew: many genes of small effect, accounting for as much as .6 or so of narrow sense heritability. However, the technique is novel and its power will improve with larger sample sizes. The small minority of skeptics who doubt the validity of twin and adoption studies now have another kind of evidence to contend with.
Our results unequivocally confirm that a substantial proportion of
ReplyDeleteindividual differences in human intelligence is due to genetic
variation, and are consistent with many genes of small effects
underlying the additive genetic influences on intelligence.
The guns, they are a-blazin'. Is this the BGI results report?
On another, maybe inappropriate note, I would like to be an author named 'A. Pickles'.
It's not our work. But I had a prepublication copy of the paper.
ReplyDeletePlomin's GWAS failed to replicate. Why should we expect this is better? (and weren't there other unreplicated GWAS for IQ?) The coverage asserting that this is a new technique is very misleading. (and if this is a new technique that we should trust more, it should say so in the abstract)
ReplyDeleteMaybe I don't understand the abstract, but isn't it saying that only 1% replicated, not the 40-50% claimed? (What does "just" SNP data mean? the only other data in the abstract is phenotype data.) I think that's the first ever replication of genetic signature of IQ. That's news. But it's only 1%. Plus, the p-value sounds wrong, but I'll have to look at the paper for that.
You are missing the distinction between a global fit of total variance accounted for and the identification of a specific locus as iq associated. This paper is not claiming any (specific gene) hits so there are none to be replicated. The overall result for total variance accounted for by SNPs (lower bound on narrow sense heritability) needs to be replicated, of course. But it's a familiar story now because of the height results.
ReplyDeleteSee my earlier post on this and the even earlier post on height. Eventually you will find visscher's paper which explains the statistical technique. You are not the only one who is confused. The genomics community is still obsessed with "missing heritability" because they don't understand the distinction above. It's yet another example of bounded cognition.
Steve,
ReplyDeleteWhat evidence is the opposition exactly resting on? Of course, I realize that genetic explanations for intelligence differences (or intelligence in general) are still not part of mainstream thinking, but I attribute that to disinterest of the general public and stupidity/ideology/cowardice among journalists. There are, however, serious thinkers who believe intelligence to not be genetic and I really want to know what their arguments are. What makes them so sure that they're right?
(When I write 'genetic', I mean gene-environment interacting factors.)
Not to beat on cosma, but you might have a look at his oft-cited blog post on iq and heritability. You will find further links there.
ReplyDelete"...3511 unrelated adults..."
ReplyDeleteBut they are very likely related in the sense that they are contemporaries and live near one another.
"We conducted a genome-wide analysis of 3511 unrelated adults with data on 549 692 single nucleotide polymorphisms (SNPs)"
Excuse my ignorance but how can just these 549 692 SNPs account for so much? Oo! Oo! The answer is "We estimate"?
"When I write 'genetic', I mean gene-environment interacting factors."
ReplyDeleteThat's the crux. The skeptics' most valid question is: What is the range of environments really?
Suppose that the Colorado twins' environments were as variable as the general population's, that would still be the general population of Colorado at a particular time.
Is population stratification a problem for this type of study?
ReplyDeleteThe skeptics have a good point there.. but even knowing that genes matter more than environment within Lake Wobegon is a fascinating/important discovery IMO.
ReplyDeleteMost people don't even accept that within their middle class environment most of the variation is due to genes.
I wonder if the method of estimating genetic distance works when the study participants are from different ethnicities.
ReplyDeleteWhen trying to determine the relationship between IQ and cranial circumference, for example, it's important to compare like with like.
The bitterenders could ask whether or not the correlation between IQ and genetic distance weren't the result of the great great ... grandchildren of the "conquerors" having it better than the great great...grandchildren of the "conquered".
Right, this is called population stratification and is one of the main problems for the genome studies. They control for it either by doing within-family comparisons or by using some statistical methods for controlling for population substructure.
ReplyDeleteThey didn't compare sibs in this study, I'm not sure whether they dealt with pop strat in some other way (article is behind a paywall).
I actually emailed Cosma about the Visscher study when it came out. This is what he said:
ReplyDelete"they're making really very drastic assumptions about the
contribution of non-genetic family effects (equal variance across the
population, uncorrelated with genetics or indeed anything else, no
interaction with genetics, etc.), which were, of course, not checked."
Sure, some more complicated effect could mimic the signal from heritability. But this study is going to be replicated using different populations from different places and eras.
ReplyDeleteIn this paper the main population was from the UK, born in the early to mid 20th century and the replication population was from Norway, aged 18-79.
If a similar study done on people from very different places and times yields a similar value for the narrow sense heritability, that would strongly suggest that the straightforward interpretation is correct.
This is a complicated topic, so more studies are necessary.
"...and the replication population was from Norway, aged 18-79..."
ReplyDeleteThe question is what is heritability for the pooled population.
"Sure, some more complicated effect could mimic the signal from heritability."
ReplyDeleteIf gene-environment interactions or correlations distort the heritability signal, replication in multiple populations won't necessarily address that.. You would need to do a study which truly controls for population substructure (e.g. family comparisons), otherwise you risk making the same sytematic error in all your samples.
Steve has shown no understanding of the need to pool populations. He has shown no appreciation of the meaninglessness of replication in another population.
ReplyDeleteThe point of studying multiple populations is to pool them. BUT they'll need to be indistinguishable genetically and distinguishable culturally for such pooled results to be meaningful.
One might compare Italians and Italian Americans or Italian Argentines. One might compare Germans and German Namibians. Only their IQs relative to their own poulation matters.
You could control for shared environmental factors by comparing siblings.
ReplyDeleteYes and check for interactions with no twins needed.
ReplyDeleteProblem with just checking for specific interactions is that the relevant environmental factors are not nearly as easy to measure as are individual genes. E.g. we might think that our crude measures of SES would capture the relevant environment, but that's not necessarily the case. Maybe the meaningful environment is made up of lots of small chaotic effects. See Turkheimer's papers "Mobiles"/"Laws of Behavior Genetics" for more on this..
ReplyDeleteMaybe I read you wrong and you agree that sib controls are necessary to address shared environmental influences.. in which case, cheers!
There's no way to unbake a cake. But it can be asked: Do the contestants end up in the same order in every bakeoff? As far as I can tell Steve does not understand this.
ReplyDeleteSure, sib controls are fine but adding other variables would let you pin point the exact effects if they exist. Some variables (e.g. SES) are hard to measure while others (divorce, number of siblings) are not. This technique would let you look for differences in these variables while controlling for genes and you could do this with a large probability sample of the population (e.g., the NLSY). This would be faster and cheaper than a twin studies. The effects may be small and chaotic but now you could prove it.
ReplyDeleteSteve,
ReplyDeleteThanks. I will.