Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence

In the paper

“Intelligence has been associated with a wide variety of human behaviors 15 and brain anatomy 27”

What the sources say

• Supported

  Citation 27 is attached to the brain-anatomy part of the sentence. In “Seguing into neuroscience,” the paper reports a positive correlation between general intelligence and brain size; “Shared genetic influences between brain structure/function and intelligence?” also describes associations with multiple brain structures. The supplied passages do not give page numbers. (LLM, conf=0.92)

  cites [15] Deary, I. J., Penke, L. & Johnson, W. The neuroscience of human intelligence differences. Nat. Rev. Neurosci. 11, 201–211 (2010). source
• Supported

  The cited paper supports the brain-anatomy part of the sentence, to which citation 27 is attached. In its main text, it reports significant correlations between general intelligence and both gray- and white-matter volume. The supplied passages do not identify a page number. (LLM, conf=0.93)

  cites [27] Posthuma, D. et al. The association between brain volume and intelligence is of genetic origin. Nat. Neurosci. 5, 83–84 (2002). source

In the paper

“Recent genome-wide metaanalyses have identified 24 genomic loci linked to variation in intelligence 3-7”

What the sources say

• Partly supported

  The Results report 13 genome-wide significant SNPs in three genomic regions associated with general cognitive function. This supports the claim that a genome-wide meta-analysis identified linked regions, but this paper does not establish the cited sentence’s total of 24 loci across multiple analyses. (LLM, conf=0.90)

  cites [3] Davies, G. et al. Genetic contributions to variation in general cognitive function: a meta-analysis of genome-wide association studies in the CHARGE consortium (N = 53949). Mol. Psychiatry 20, 183–192 (2015). sourceExperimental — accuracy not yet measured
• Partly supported

  The Introduction (page unavailable) reports genome-wide significant associations with general cognitive function, and the Results section on verbal–numerical reasoning (page unavailable) identifies three associated genomic regions. These passages support the link between genomic loci and cognitive variation, but the paper does not establish the claimed total of 24 loci identified by recent meta-analyses. (LLM, conf=0.85)

  cites [4] Davies, G. et al. Genome-wide association study of cognitive functions and educational attainment in UK Biobank (N = 112 151). Mol. Psychiatry 21, 758–767 (2016). sourceExperimental — accuracy not yet measured
• Partly supported

  The Letters section reports that this genome-wide meta-analysis identified loci associated with intelligence, supporting the general point. It reports 18 independent significant loci, not 24; the cited paper alone therefore does not substantiate the sentence’s total across multiple studies. No page number is supplied in our copy. (LLM, conf=0.88)

  cites [5] Sniekers, S. et al. Genome-wide association meta-analysis of 78,308 individuals identifies new loci and genes influencing human intelligence. Nat. Genet. 49, 1107–1112 (2017). sourceExperimental — accuracy not yet measured
• Partly supported

  The Results (GWAS of general cognitive function; page unavailable) report two genome-wide-significant loci, supporting the claim that a meta-analysis identified loci associated with cognitive performance. The Abstract reports that 24 is the number of cohorts studied, not the number of loci identified; it also reports three additional loci from gene-based analysis. This source therefore does not support the count of 24 genomic loci. (LLM, conf=0.90)

  cites [6] Trampush, J. W. et al. GWAS meta-analysis reveals novel loci and genetic correlates for general cognitive function: a report from the COGENT consortium. Mol. Psychiatry 22, 336–345 (2017). sourceExperimental — accuracy not yet measured
• Partly supported

  The Results (Analysis of genome-wide SNPs; page unavailable) identify three significant SNPs at one ADAM12 locus, and the Introduction discusses earlier genome-wide studies of intelligence. Neither passage supports the specific claim that meta-analyses identified 24 loci. (LLM, conf=0.83)

  cites [7] Zabaneh, D. et al. A genome-wide association study for extremely high intelligence. Mol. Psychiatry 23, 1226–1232 (2018). sourceExperimental — accuracy not yet measured

In the paper

“a major determinant of human health and well-being 2”

What the sources say

• Partly supported

  The Abstract’s Results & conclusion and the Discussion (page numbers not provided) link higher intelligence in youth with better physical health and fewer chronic conditions at age 50. They describe associations, not intelligence as a major determinant of human health generally, and the study does not establish an effect on well-being. (LLM, conf=0.85)

  cites [2] Wraw, C., Deary, I. J., Gale, C. R. & Der, G. Intelligence in youth and health at age 50. Intelligence 53, 23–32 (2015). sourceExperimental — accuracy not yet measured

In the paper

“Such findings are consistent with previous studies implicating bidirectional causal effects 30,31”

What the sources say

• Supported

  The paper’s discussion of prior research says studies indicate that intelligence has causal effects on education. It also reports that studies of schooling generally found more time in school leads to greater intelligence. The claim’s word “implicating” appropriately reflects the paper’s caution that causal effects may flow in both directions (page unavailable in our copy). (LLM, conf=0.91)

  cites [30] Johnson, W., Deary, I. J. & Iacono, W. G. Genetic and environmental transactions underlying educational attainment. Intelligence 37, 466–478 (2009). source
• ?Not checked — no full text of the sourcecites [31] Richards, M. & Sacker, A. Is education causal? Yes. Int. J. Epidemiol. 40, 516–518 (2011). source

In the paper

“Conserved and regulatory regions have previously been implicated in cognitive functioning 22 , but coding regions have not.”

What the sources say

• Partly supported

  The Results (Partitioned heritability; page unavailable) report enrichment for cognitive-function heritability in conserved regions and in regulatory annotations, including weak enhancers and DNase hypersensitivity sites. The paper does not establish the broader claim that coding regions have not been implicated in cognitive functioning. (LLM, conf=0.82)

  cites [22] Hill, W. D. et al. Molecular genetic aetiology of general cognitive function is enriched in evolutionarily conserved regions. Transl. Psychiatry 6, e980 (2016). sourceExperimental — accuracy not yet measured

In the paper

“A number of previous reports support both a causal link and genetic overlap between these phenotypes 32,33”

What the sources say

• Supported

  The Discussion reports prior evidence of genetic overlap between IQ and schizophrenia liability, including a twin study finding a strong negative genetic correlation. Its Conclusions say the IQ–schizophrenia association may be causal, while clarifying that shared genetic risk does not explain the association in this sample. (LLM, conf=0.88)

  cites [32] Kendler, K. S., Ohlsson, H., Sundquist, J. & Sundquist, K. IQ and schizophrenia in a Swedish national sample: their causal relationship and the interaction of IQ with genetic risk. Am. J. Psychiatry 172, 259–265 (2015). source
• Partly supported

  The Discussion and the ‘Genetic Overlap of SCZ and Educational Attainment’ section report genetic overlap between general cognition and schizophrenia; page numbers are unavailable in our copy. The paper does not establish a causal link between intelligence and schizophrenia, so it supports only the genetic-overlap part of the claim. (LLM, conf=0.88)

  cites [33] Le Hellard, S. et al. Identification of gene loci that overlap between schizophrenia and educational attainment. Schizophr. Bull. 43, 654–664 (2017). sourceExperimental — accuracy not yet measured

In the paper

“We used an FDR of 1 × 10 -5 to define significant interactions, based on previous recommendations 51 modified to account for the differences in cell lines used here.”

What the sources say

• Partly supported

  The Experimental Procedures ('Identification of significant Hi-C contacts') support using an FDR threshold to define significant contacts, but specify FDR < 1 × 10⁻⁶, not 1 × 10⁻⁵. The Results ('Identification of Frequently Interacting Regions') discuss variation between samples, but do not describe the citing authors’ cell-line-specific modification. (LLM, conf=0.84)

  cites [51] Schmitt, A. D. et al. A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Rep. 17, 2042–2059 (2016). sourceExperimental — accuracy not yet measured

In the paper

“Briefly, brain-cell-type expression data were drawn from single-cell RNA-seq data from mouse brains.”

What the sources say

• Supported

  In Results, under “Cell type specificity of gene expression,” the paper describes assembling brain single-cell RNA-seq data and identifies the resulting dataset as the “KI mouse superset.” It then uses those data to estimate gene-expression specificity by cell type, supporting the claim about the source of the cell-type expression data. (LLM, conf=0.93)

  cites [53] Skene, N. G. et al. Genetic identification of brain cell types underlying schizophrenia. Nat. Genet. 50, 825–833 (2018). source
• Supported

  The Materials and Methods section (‘Tissue- and cell-specific gene expression’) explicitly says brain cell-type expression data came from scRNA-seq data from mouse brain. It also describes calculating expression scores for each cell type, consistent with using those data to construct the cell-type gene sets. (LLM, conf=0.95)

  cites [37] Coleman, J.R.I. et al. Biological annotation of genetic loci associated with intelligence in a meta-analysis of 87,740 individuals. Mol. Psychiatry https://doi.org/10.1038/s41380-018-0040-6 (2018). source

In the paper

“Comparison with previous GWAS 28 supported these correlations, showing numerous shared genetic variants across phenotypes”

What the sources say

• Partly supported

  The Introduction (page not provided) says GWAS Catalog data have been used to find evidence of pleiotropy, which supports the general idea of genetic associations across traits. It does not report the citing authors’ comparison with previous GWAS, confirm their correlations, or show that numerous variants were shared across the phenotypes they studied. (LLM, conf=0.80)

  cites [28] MacArthur, J. et al. The new NHGRI-EBI Catalog of published genome-wide association studies (GWAS Catalog). Nucleic Acids Res. 45(D1), D896–D901 (2017). sourceExperimental — accuracy not yet measured

In the paper

“A CADD score above 12.37 is the threshold to be potentially pathogenic 23 .”

What the sources say

• Partly supported

  The Technical Report says CADD scores rank known pathogenic variants highly, but it does not establish 12.37 as a pathogenicity threshold. It also says C-scores cannot be precisely calibrated to the likelihood that a variant is pathogenic and are best interpreted as indicating likelihood of deleteriousness. (LLM, conf=0.87)

  cites [23] Kircher, M. et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nat. Genet. 46, 310–315 (2014). sourceExperimental — accuracy not yet measured

In the paper

“The estimated causal effect coefficients (b xy ) are approximately equal to the natural log odds ratio (OR) for a case-control trait 29 .”

What the sources say

• Supported

  The Results (Overview of the methods) state that the GSMR estimate of bxy approximately equals logOR, where OR is the effect of a risk factor on disease without confounding. The Methods (The GSMR method) add that the approach applies to case–control disease data because logOR is free of ascertainment bias. (LLM, conf=0.94)

  cites [29] Zhu, Z. et al. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat. Commun. 9, 224 (2018). source

In the paper

“This method improves power when using an extreme case sampling design such as that in HiQ 35 and provides a comparable metric with which to combine information from different analytic designs while accounting for their differences in power/effective sample size.”

What the sources say

• Supported

  Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites [35] Peloso, G. M. et al. Phenotypic extremes in rare variant study designs. Eur. J. Hum. Genet. 24, 924–930 (2016). source

In the paper

“Independently associated loci from the metaanalysis were defined using FUMA 24 (see URLs), an online platform for functional mapping of genetic variants.”

What the sources say

• Supported

  The Results section (‘Overview of FUMA web application’) says FUMA identifies independent significant SNPs and their surrounding genomic loci from supplied GWAS summary statistics. The Methods section (‘Characterization of genomic risk loci based on GWAS’) describes how it defines those loci. This supports citing FUMA as the platform used for that step; the source does not independently document the citing authors’ meta-analysis. (LLM, conf=0.93)

  cites [24] Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nat. Commun. 8, 1826 (2017). source

In the paper

“Convergent evidence of strong association (z = 9.49) and the highest observed probability of a deleterious protein effect (CADD 23 score = 34) were found for rs13107325.”

What the sources say

• Supported

  The citation marker appears immediately after “CADD,” identifying the scoring method rather than sourcing the rs13107325 result. The Technical Report (page unavailable) describes CADD as a framework for scoring human variants by estimated deleteriousness; the z statistic and score of 34 are not the part attributed to this source. (LLM, conf=0.92)

  cites [23] Kircher, M. et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nat. Genet. 46, 310–315 (2014). source

In the paper

“This method can help differentiate the likely causal direction of association between two traits but cannot make any statement about the intermediate mechanisms involved in any potential causal process.”

What the sources say

• Partly supported

  The Methods section describes GSMR as testing whether an exposure causally affects an outcome, and the Results section uses forward and reverse GSMR analyses to assess direction. The Results also say it is difficult to distinguish whether an effect is mediated by other risk factors or driven by them. But the paper does examine associations among risk factors and conditional effects, so the categorical claim that GSMR cannot make any statement about intermediate mechanisms overstates its stated limitation. Page numbers are unavailable in our copy. (LLM, conf=0.86)

  cites [29] Zhu, Z. et al. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat. Commun. 9, 224 (2018). sourceExperimental — accuracy not yet measured

In the paper

“FUMA currently contains Hi-C data for 14 tissue types from the study of Schmitt et al. 51 .”

What the sources say

• Supported

  The Results section, “A compendium of chromatin organization maps across 21 human cell and tissue types,” states that Schmitt et al. conducted Hi-C analysis on 14 primary human tissues. This supports the citation identifying the study as the source of those data; it does not independently verify FUMA’s contents. (LLM, conf=0.92)

  cites [51] Schmitt, A. D. et al. A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Rep. 17, 2042–2059 (2016). source

In the paper

“comparable heritability across age groups, as captured by the SNPs included in the analysis (h SNP 2 = 0.19-0.22)”

In the paper

“Conditional analysis indicated that there were three independent associations-regulation of nervous system development, central nervous system neuron differentiation, and regulation of synapse structure or activity-that together accounted for the associations of the other sets.”

In the paper

“of which 350 were also mapped by FUMA”

In the paper

“The FUMA-mapped genes were enriched for brain tissue expression and several regulatory biological gene sets”

In the paper

“Cross-locus interactions implicated ELAVL2, PTCH1, ATF4, FBXL17, and MAN2A1 in the left ventricle of the heart, SATB2 in liver tissue, and MEF2C in five tissues.”

In the paper

“This missense mutation (MAF = 0.065, P = 2.23 × 10 -21 ) in SLC39A8 was the lead SNP in locus 71”

In the paper

“twin and family studies show strong genetic correlations across diverse cognitive domains 11 , suggesting pleiotropy, and across levels of ability 11”

What the sources say

• ?Not checked — no full text of the sourcecites [11] Plomin, R. & Kovas, Y. Generalist genes and learning disabilities. Psychol. Bull. 131, 592–617 (2005). source

In the paper

“the variance common across cognitive tasks can be modeled as a latent factor denoted as g (the general factor of intelligence) 9,10”

What the sources say

• ?Not checked — no full text of the sourcecites [9] Carroll, J. B. Human Cognitive Abilities: A Survey of Factor-Analytic Studies. (Cambridge University Press, Cambridge, UK, 1993). source
• ?Not checked — no full text of the sourcecites [10] Spearman, C. “General intelligence,” objectively determined and measured. Am. J. Psychol. 15, 201–292 (1904). source

In the paper

“The Bonferroni-corrected significance threshold was 0.05/56 annotations = 0.0009.”

In the paper

“This reduced the number of genes mapped but increased the likelihood that those identified would have a plausible biological function.”

In the paper

“SNP-based replication was carried out using a proxy phenotype of educational attainment in an independent sample.”

In the paper

“We used a threshold P value of 0.01 for the outlier detection analysis in HEIDI, which removes 1% of SNPs by chance if there is no pleiotropic effect.”

In the paper

“With the exception of HiQ/HRS, all cohorts extracted a single sum score, mean score, or factor score from a multidimensional set of cognitive performance tests and used this normally distributed score as the phenotype in a covariate-adjusted (for example, age, sex, ancestry principal components) GWAS using linear regression methods.”

In the paper

“The Bonferroni-corrected significance threshold was 0.05/38 traits = 1.32 × 10 -3 .”

In the paper

“Many of these have been identified by non-GWAS sources and are not represented in the NHGRI catalog.”

In the paper

“We calculated polygenic scores (PGSs) based on the SNP effect sizes of the leave-one-out meta-analyses, from which four cohorts were (separately) excluded and reserved for score validation.”

In the paper

“The X chromosome was treated separately in the meta-analysis because imputed genotypes were not available for the X chromosome in the largest cohort (UKB), and there was little overlap between the UKB called genotypes and imputed data from other cohorts (n SNPs < 500).”

In the paper

“We therefore included only the called X-chromosome variants in UKB for these analyses after performing X-chromosome-specific quality-control steps 38 .”

What the sources say

• ?Not checked — no full text of the sourcecites [38] König, I. R., Loley, C., Erdmann, J. & Ziegler, A. How to include chromosome X in your genome-wide association study. Genet. Epidemiol. 38, 97–103 (2014). source

In the paper

“The Bonferroni-corrected significance threshold was 0.05/9,494 gene sets = 5.27 × 10 -6 .”

In the paper

“We used a false discovery rate (FDR) of 0.05 to define significant eQTL associations”

In the paper

“The explained variance (Δ R 2 ) was derived from a linear model in which the GWAS intelligence phenotype was regressed on each PGS while controlling for the same covariates as in each cohort-specific GWAS, compared to a linear model with GWAS covariates only.”

In the paper

“All files were checked for data integrity and accuracy.”

In the paper

“The Bonferroni-corrected significance threshold was 0.05/7,323 gene sets = 6.83 × 10 -6 .”

In the paper

“After combining all data, meta-analysis results were further filtered to exclude any variants with n < 50,000.”

In the paper

“We first identified 'independent significant SNPs' , which had a Bonferroni-corrected genome-wide significant two-tailed P value (P < 5 × 10 -8 ) and represented signals that were independent from each other at r 2 < 0.6.”

In the paper

“Indels and SNPs that were duplicated, multiallelic, monomorphic, or ambiguous (A/T or C/G with MAF > 0.4) were also excluded.”

In the paper

“The Bonferroni-corrected genome-wide significance threshold was 0.05/18,128 genes = 2.76 × 10 -6 .”

In the paper

“To create independent discovery samples for use in polygenic score validation, we also conducted meta-analyses with a 'leave-one-out' strategy in which summary statistics from four validation datasets were each excluded from the meta-analysis”

In the paper

“SNPs were filtered from further analysis if they met any of the following criteria: imputation quality (INFO/R 2 ) score < 0.6, Hardy-Weinberg equilibrium P < 5 × 10 -6 , study-specific minor allele frequency (MAF) corresponding to a minor allele count (MAC) < 100, and mismatch of alleles or allele frequency difference greater than 20% from the Haplotype Reference Consortium (HRC) genome reference”

What the sources say

• ?Not checked — no full text of the sourcecites [16] Deary, I. J. Intelligence. Annu. Rev. Psychol. 63, 453–482 (2012). source

In the paper

“Visual inspection of the distribution of the summary statistics was performed, and Manhattan plots and quantile-quantile plots were created for the cleaned summary statistics from each cohort (Supplementary Fig. 1).”

In the paper

“Because chromatin interactions are often defined in a certain resolution, such as 40 kb, an interacting region can span multiple genes.”

In the paper

“Participants ranged from children to older adults, with older samples being screened for cognitive decline to exclude the possibility of dementia affecting performance on cognitive tests.”

In the paper

“We additionally included a random-effects meta-analysis for each phase, as implemented in METAL, to evaluate potential heterogeneity in the SNP association statistics between cohorts.”

In the paper

“For each independent genomic locus, we considered it to be evidence for replication if the lead SNP or another correlated SNP in the region was sign concordant with the corresponding SNP in the intelligence meta-analysis and had a two-tailed P value of association with educational attainment smaller than 0.05/242 independent tests = 0.0002.”

In the paper

“To test replication of the SNPs with this proxy phenotype, we performed a sign concordance test for all genome-wide significant SNPs from the meta-analysis using the two-tailed exact binomial test.”