The Unz Review: An Alternative Media Selection
A Collection of Interesting, Important, and Controversial Perspectives Largely Excluded from the American Mainstream Media
 TeasersGene Expression Blog
Toward Discovering Genes Affecting Cognition
🔊 Listen RSS
Email This Page to Someone

 Remember My Information


Bookmark Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New ReplyRead More
ReplyAgree/Disagree/Etc. More... This Commenter This Thread Hide Thread Display All Comments
These buttons register your public Agreement, Disagreement, Troll, or LOL with the selected comment. They are ONLY available to recent, frequent commenters who have saved their Name+Email using the 'Remember My Information' checkbox, and may also ONLY be used once per hour.
Ignore Commenter Follow Commenter
Search Text Case Sensitive  Exact Words  Include Comments
List of Bookmarks

Citation: Common genetic variants influence human subcortical brain structures, Nature (2015) doi:10.1038/nature14101

Citation: Common genetic variants influence human subcortical brain structures, Nature (2015) doi:10.1038/nature14101

Here’s what we know. Intelligence, as defined by a general factor which explains variation across a range of cognitive tasks, is substantially heritable, with a narrow sense heritability on the order of 0.25 to 0.75 depending on who you talk to and what context.* Intelligence itself exhibits correlations with other traits, from those of social importance, such as education, as well as biological parameters, such as brain size. Additionally, the effect size of genetic variants associated with general intelligence are likely to be very small. This means that you should be immediately skeptical of claims that a common variant segregating in the population explains a large proportion of the variation in intelligence within the population. The history of this area of research, which goes back to linkage studies, is one of non-reproducibility. Large effect quantitative trait loci should already have been picked up by linkage studies decades ago, so I am usually rather skeptical when this old wine is presented again in a genomic guise. In short, the genetic architecture of general intelligence is likely to resemble height, with many loci of small effect.**

This is what Rietveld et al. found last fall in Common genetic variants associated with cognitive performance identified using the proxy-phenotype method. The same sizes were on the order of 10,000 to 100,000 within this study. The top associations within this study explain less than 1% of the variation within the data. It seems likely that the largest effect alleles which influence intelligence variation are about an order of magnitude smaller in impact than those for height. A new paper in Nature, Common genetic variants influence human subcortical brain structures, looks at the morphology of the brain, synthesizing imaging, cognitive neuroscience, and genomics. Here’s the abstract:

…To investigate how common genetic variants affect the structure of these brain regions, here we conduct genome-wide association studies of the volumes of seven subcortical regions and the intracranial volume derived from magnetic resonance images of 30,717 individuals from 50 cohorts. We identify five novel genetic variants influencing the volumes of the putamen and caudate nucleus. We also find stronger evidence for three loci with previously established influences on hippocampal volume and intracranial volume. These variants show specific volumetric effects on brain structures rather than global effects across structures. The strongest effects were found for the putamen, where a novel intergenic locus with replicable influence on volume (rs945270; P = 1.08 × 10−33; 0.52% variance explained) showed evidence of altering the expression of the KTN1 gene in both brain and blood tissue. Variants influencing putamen volume clustered near developmental genes that regulate apoptosis, axon guidance and vesicle transport. Identification of these genetic variants provides insight into the causes of variability in human brain development, and may help to determine mechanisms of neuropsychiatric dysfunction.

Paul Thompson was involved in the research, so I am confident that it was be done thoroughly (and the author list is long enough that I hope they checked for obvious problems!). To correct for population stratification within this European sample they looked at the top for dimensions of variation, and used a regression model to capture other variables which might be confounded with the SNPs in question. The small proportion of variation explained actually increases my confidence, in that it seems to be in the same order of magnitude as the type of studies looking at endophenotypes.

Because of their sheer number I doubt that there’s a great short term likelihood of annotating all the genes responsible for variation in intelligence. Rather, I wonder if the ultimate goal is something similar to what occurred with statins. Find a small effect locus, and target a drug at that locus to help cure cognitive illnesses such as schizophrenia. It stands to reason that the same loci which impact general intelligence would also shape cognitive phenotypes which we term pathological.

* So if heritability in the narrow sense is 0.50 that means half the variation in intelligence in the population can be explained by variation of genes in the population. By way of comparison, height is 0.80 to 0.90 heritable in the narrow sense in the developed world. This does not mean that the correlation between parents and offspring is 0.80 or 0.90 for height. In fact the correlation is closer to 0.50 for height between parents and offspring and also between siblings.

** An alternative minority viewpoint is many rare alleles of somewhat larger effect.

• Category: Science • Tags: Genomics, GWAS, IQ 
Hide 10 CommentsLeave a Comment
Commenters to FollowEndorsed Only
Trim Comments?
  1. Isn’t the advantage of the “alternative view about numerous very rare alleles with stronger effects” in the possibility of these stronger affects being largely detrimental / the variants being deleterious? It’s harder to imagine persistence of common deleterious alleles when it comes to cognition (and higher-frequency alleles ought to be old). In contrast, recently-emerged and therefore rare deleterious alleles could be continuously lost to selection, never reaching high frequencies?

  2. Your point about small-effect variants being possible drug targets is very true, but it is worth pointing out that in the case of statins the critical role of HMG-CoA reductase in cholesterol synthesis was very well known, and it was a very hotly anticipated drug target. It is possible that genetic studies will unveil a variant of small effect that turns out to be a crucial component of some pathway very central to the development of schizophrenia, but in our current state of knowledge, we have no idea if any such pathway even exists.
    I am not saying it’s not possible, just mentioning this in case someone thinks statins came to us via a small effect genetic discovery.. (I know you don’t think that).

  3. I thought the idea of a general factor was falling increasingly out of favour, with it really being a proxy for multiple “types” of intelligence. Or am I behind on my reading there?

  4. #1, i think some studies have already shown this is the case for de novo mutations. though i don’t think that the QTLs are going to be THAT large. else they would have shown up in family based linkage studies. most people don’t seem to buy goldstein’s idea of ‘synthetic mutations’.

    #2, good point. though do note that eric lander uses the argument i implied pretty explicitly.

    #3, there are still arguments about this. most people don’t care about psychometrics, so i think that most people outside of it accept ‘multiple intelligences,’ but mostly on ideological grounds since IQ/g factor is icky. without the field it’s still pretty useful.

  5. It used to worry me that twin studies consistently found high heritability of IQ, but that molecular studies could never seem to find any predictive alleles.

    Reading this paper reassured me:
    Using a model that simultaneously fits effects of 300k SNPs (with random effects) essentially reproduces the results of twin studies. Controlled for PC’s and excluded relatives. The cool thing is that the effect of each SNP is only poorly estimated (so that only a small proportion are significant, and so they only explain 1% of variance), but when you aggregate information across all of them, they confidently predict lots of variance. Cool stuff.

    Razib – there must be ways to distinguish the common-alleles-of-small-effect hypothesis from the rare-alleles-of-large-effect. Hmm?

  6. My mistake: 550k SNPs

    It occurs to me that the above paper used *common* SNPs. So it must be that common SNPs account for a large part of the heritable variance.

  7. There have been recent reports from James Sikela et al about influence on cognitive traits of allelic variation at a gene called DUF1220 which involves copy number variation (?? so maybe missed by techniques that only look at SNP variation??).

    Recently the British psychometrician James Thompson posted a query about this work on his blog, Thompson asked if a population geneticist would comment on it.

    The paper in question is:

    DUF1220 copy number is linearly associated with increased cognitive function as measured by total IQ and mathematical aptitude scores
    Jonathon M. Davis · Veronica B. Searles · Nathan Anderson ·Jonathon Keeney · Armin Raznahan · L. John Horwood · David M. Fergusson · Martin A. Kennedy · Jay Giedd · James M. Sikela

    Hum Genet (2015) 134:67–75 DOI 10.1007/s00439-014-1489-2

    So Razib, in your opinion is this more of the “old wine” sort of unreproducible gene-IQ linkage claims, or could it be valid??

  8. #7, the paper was brought to my attention. and yes, it is not something i think will stand up. effect size is just too large. just because it is a cnv, it should have shown up as an association in some linkage study if it is so strong.

    (the statistics in the paper btw do suggest this is a case of a false positive)

  9. #4 Synthetic genome-wide associations of kind they described
    are a relatively common occurrence in the tumor suppressor gene studies, which is my narrow field of expertise. It’s caused in part by the nature of the mutations (which tend to be loss-of-function deleterious, and often dominant) but also by the relatively modest pace of purifying selection (since carrying a mutation doesn’t preclude procreation). So the founder mutations may reach relatively higher frequencies, only an order to 2 orders of magnitude less common than the study SNPs found in association with them, resulting in synthetic associations even in moderate-size studies.

    If the evaluated phenotype is common, and if the penetrance isn’t extremely high, then the linkage studies are powerless to detect the genes. Too many phenocopies, too many carriers not reaching the set phenotype thresholds or weighing too little for a QTL study. Affinity mate selection patterns (which is known e.g. in obesity linkage attempts) makes matters even worse for any linkage hopes.

    The heritable deleterious variants affective cognition ought to be less penetrant than the mutations typically seen de novo. But could be in the same genes / pathways.

  10. […] the work force, the dismal academic performance, and the high dependency on welfare, much less the research on the genetics of IQ. I mean, we’re talking about a community whose highest profile issue […]

Comments are closed.

Subscribe to All Razib Khan Comments via RSS