A global overview of pleiotropy and genetic architecture in complex traits

Abstract

ince the first genome-wide association study on macular degeneration in 2005 (ref. 1), over 3,000 GWASs have been published, for over 1,000 traits, reporting on tens of thousands of genetic risk variants 2. These results have increased our understanding of the genetic architecture of traits. Occasionally, GWAS results have led to further insight into disease mechanisms 3,4 , such as autophagy for Crohn's disease 5 , immunodeficiency for rheumatoid arthritis 6 and transcriptome regulation through FOXA2 in the pancreatic islet and liver for type 2 diabetes 7. After a decade of GWASs, we have learned that the majority of studied traits are highly polygenic and influenced by many genetic variants, each of small effect 4,8 , with disparate genetic architectures across traits 9. Fundamental questions (such as whether all genetic variants or genes in the human genome are associated with at least one, many or even all traits, and whether the polygenic effects for specific traits are functionally clustered or randomly spread across the genome) are, however, still unanswered 4,10,11. Such knowledge would greatly enhance our understanding of how genetic variation leads to trait variation and trait correlations. Whereas GWAS primarily aims to discover genetic variants associated with specific traits, the current availability of vast amounts of GWAS results allow investigation of these general questions. To this end, we compiled a catalog of 4,155 GWAS results across 2,965 unique traits from 295 studies (https://atlas.ctglab.nl), including publicly available GWASs and new results for 600 traits from the UK Biobank 12. These GWAS results were used in the current study to (1) chart the extent of pleiotropy at trait-associated locus, gene, SNP and gene-set levels, (2) characterize the nature of trait-associated variants (that is, the distribution of effect size, minor allele frequency (MAF) and biological functionality of trait-associated or credible SNPs) and (3) investigate genetic architecture across a variety of traits and domains in terms of SNP heritability and trait polygenicity (see Supplementary Fig. 1). Results Catalog of 4,155 GWAS summary statistics. We collected publicly available, full GWAS summary statistics (last update 23 October 2018; see Methods) resulting in 3,555 sets of GWAS summary statistics from 294 studies. We additionally performed GWAS on 600 traits available from the UK Biobank release 2 cohort (UKB2; release May 2017) 12 , by selecting nonbinary traits with >50,000 European individuals with nonmissing phenotypes, and binary traits for which the number of available cases and controls were both >10,000 and total sample size was >50,000 (see Methods, Supplementary Note and Supplementary Tables 1 and 2). In total, we collected 4,155 GWASs from 295 unique studies covering 2,965 unique traits (Supplementary Table 3). Traits were classified into 27 domains 13,14. The average sample size across curated GWASs was 56,250 subjects, with a maximum of 898,130 for type 2 diabetes 15. The 4,155 GWAS results are ma...

Shared risk alleles with discordant polygenic effects: Disentangling the genetic overlap between ASD and ADHD

“…The pattern of ASD-and ADHD-specific associations with EA, in combination with discordant polygenic cross-disorder links, was (i) reciprocally detectable using both ASD and ADHD-related variant sets, (ii) replicated at P thr <0.05 using ASD(PGC) summary statistics, (iii) independent of risk allele alignment for ASD and ADHD and (iv) consistent with the previously reported genetic overlap between EA, ASD and ADHD 21,28 . This suggests that cross-disorder associations are driven by a substantial proportion of subthreshold variants that are associated with both ASD and ADHD, reflecting pleiotropic effects, presumed for many trait-associated variants in the genome 33 . Moreover, joint modelling of these pleiotropic alleles, exploiting single instead of aggregated SNP estimates, could substantially increase evidence for both disorder-specific and cross-disorder associations.…”

A Rare Mutation in SMAD9 Associated With High Bone Mass Identifies the SMAD-Dependent BMP Signaling Pathway as a Potential Anabolic Target for Osteoporosis

“…nl/), an online database of publicly available summary results statistics from 4,155 GWAS from 295 unique studies across 2960 unique traits and 27 domains. (19) Significance for pleiotropic associations used a traditional genome-wide significance threshold for SNP-trait PheWAS (p < 5 × 10 −8 ). (17)…”

“…SMAD9 HIGH BONE MASS 99 n Phenomewide association study PheWAS involving nearly 3000 traits (19) identified no clear evidence for pleiotropy for the c.65T>C, p.Leu22Pro SMAD9 variant (rs111748421) (Supplemental Fig. S6 and Supplemental Table S4 in File S1).…”

Mechanisms of tissue-specific genetic regulation revealed by latent factors across eQTLs

“…6F), which has not been previously shown. HNF4A is an essential TF during liver organogenesis and development [38,39] and harbors a missense mutation (rs1800961) strongly associated with liver relevant traits including high-density lipoprotein levels and total cholesterol [45,46,47] (Additional file 1: Figure S16).…”