Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 9 (11), e112390
eCollection

Results of a "GWAS Plus:" General Cognitive Ability Is Substantially Heritable and Massively Polygenic

Affiliations

Results of a "GWAS Plus:" General Cognitive Ability Is Substantially Heritable and Massively Polygenic

Robert M Kirkpatrick et al. PLoS One.

Erratum in

Abstract

We carried out a genome-wide association study (GWAS) for general cognitive ability (GCA) plus three other analyses of GWAS data that aggregate the effects of multiple single-nucleotide polymorphisms (SNPs) in various ways. Our multigenerational sample comprised 7,100 Caucasian participants, drawn from two longitudinal family studies, who had been assessed with an age-appropriate IQ test and had provided DNA samples passing quality screens. We conducted the GWAS across ∼ 2.5 million SNPs (both typed and imputed), using a generalized least-squares method appropriate for the different family structures present in our sample, and subsequently conducted gene-based association tests. We also conducted polygenic prediction analyses under five-fold cross-validation, using two different schemes of weighting SNPs. Using parametric bootstrapping, we assessed the performance of this prediction procedure under the null. Finally, we estimated the proportion of variance attributable to all genotyped SNPs as random effects with software GCTA. The study is limited chiefly by its power to detect realistic single-SNP or single-gene effects, none of which reached genome-wide significance, though some genomic inflation was evident from the GWAS. Unit SNP weights performed about as well as least-squares regression weights under cross-validation, but the performance of both increased as more SNPs were included in calculating the polygenic score. Estimates from GCTA were 35% of phenotypic variance at the recommended biological-relatedness ceiling. Taken together, our results concur with other recent studies: they support a substantial heritability of GCA, arising from a very large number of causal SNPs, each of very small effect. We place our study in the context of the literature-both contemporary and historical-and provide accessible explication of our statistical methods.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1
Figure 1. Manhattan plot of GWAS p-values, all 2,546,647 observed and imputed SNPs.
Chromosome 23 = X chromosome, chromosome 25 = pseudoautosomal region of sex chromosome. Chromosome 26 indicates mitochondrial DNA. SNPs are plotted by serial position on each chromosome. Genome-wide significance is –log10(p) >7.30, which no SNP reaches. The peak on chromosome 1 is due to 11 SNPs (rs10922924, rs3856228, plus 9 others imputed nearby) that span about 14 kb not within a known gene. The peaks on chromosomes 16 and 21 are each due to a single imputed SNP, respectively rs16947526 and rs9982370.
Figure 2
Figure 2. Uniform quantile-quantile plot of GWAS p-values, all 2,546,647 observed and imputed SNPs.
The black curves delineate 95% confidence limits.
Figure 3
Figure 3. Manhattan plot for gene-based p-values from VEGAS.
Analysis input was GWAS p-values from 2,485,149 autosomal SNPs, both observed and imputed. Abscissa position of each point is the gene’s beginning base-pair position, NCBI genome build 36. Genome-wide significance is –log10(p) >5.55, which no gene reaches.
Figure 4
Figure 4. Uniform quantile-quantile plot for gene-based p-values from VEGAS.
Analysis input was GWAS p-values from 2,485,149 autosomal SNPs, both observed and imputed. Black curves delineate 95% confidence limits.
Figure 5
Figure 5. Five-fold cross-validation R 2 of polygenic score (averaged across subsamples, in black) for predicting FSIQ residualized for covariates, compared to results from simulated null data (98th percentiles, in red).
Black lines represent cross-validation Buse’s R 2 for predicting residualized FSIQ, averaged across the 5 subsamples. In each subsample, residualized FSIQ was predicted from polygenic score calculated from regression weights obtained in the other 4 subsamples. “P-value cutoff” dictated how small a SNP’s p-value had to be in the calibration GWAS to be included in calculating polygenic score for the validation sample. Red lines represent the results of 50 iterations of parametric bootstrapping, which conducted polygenic scoring under cross-validation using data simulated under the null of independence between phenotype and SNP genotypes (conditional on covariates). Each point plotted for the red lines is the 98th percentile, among the 50 iterations of parametric bootstrapping, of R 2 at that p-value cutoff. Polygenic score was either calculated directly from the GWAS weights (solid lines) or from signed unit weights (dashed lines; see text).

Similar articles

See all similar articles

Cited by 18 articles

See all "Cited by" articles

References

    1. Gottfredson LS (2003) g, Jobs, and Life. In: Nyborg H, editor. The Scientific Study of General Intelligence: Tribute to Arthur R. Jensen. New York: Pergamon. 293–342.
    1. Herrnstein RJ, Murray C (1994) The Bell Curve: Intelligence and Class Structure in American Life. New York: Simon & Schuster, Inc.
    1. Jensen AR (1998) The g Factor: The Science of Mental Ability. London: Praeger.
    1. Deary IJ (2012) Intelligence. Annual Review of Psychology 63: 453–482 10.1146/annurev-psych-120710-100353 - DOI - PubMed
    1. Bouchard TJ, McGue M (1981) Familial studies of intelligence: A review. Science 212: 1055–1059. - PubMed

Publication types

Feedback