Comprehensive evaluation of disease- and trait-specific enrichment for eight functional elements among GWAS-identified variants

Hum Genet. 2017 Jul;136(7):911-919. doi: 10.1007/s00439-017-1815-6. Epub 2017 May 31.

Abstract

Genome-wide association study (GWAS)-identified variants are enriched for functional elements. However, we have limited knowledge of how functional enrichment may differ by disease/trait and tissue type. We tested a broad set of eight functional elements for enrichment among GWAS-identified SNPs (p < 5×10-8) from the NHGRI-EBI Catalog across seven disease/trait categories: cancer, cardiovascular disease, diabetes, autoimmune disease, psychiatric disease, neurological disease, and anthropometric traits. SNPs were annotated using HaploReg for the eight functional elements across any tissue: DNase sites, expression quantitative trait loci (eQTL), sequence conservation, enhancers, promoters, missense variants, sequence motifs, and protein binding sites. In addition, tissue-specific annotations were considered for brain vs. blood. Disease/trait SNPs were compared to a control set of 4809 SNPs matched to the GWAS SNPs (N = 1639) on allele frequency, gene density, distance to nearest gene, and linkage disequilibrium at ~3:1 ratio. Enrichment analyses were conducted using logistic regression, with Bonferroni correction. Overall, a significant enrichment was observed for all functional elements, except sequence motifs. Missense SNPs showed the strongest magnitude of enrichment. eQTLs were the only functional element significantly enriched across all diseases/traits. Magnitudes of enrichment were generally similar across diseases/traits, where enrichment was statistically significant. Blood vs. brain tissue effects on enrichment were dependent on disease/trait and functional element (e.g., cardiovascular disease: eQTLs P TissueDifference = 1.28 × 10-6 vs. enhancers P TissueDifference = 0.94). Identifying disease/trait-relevant functional elements and tissue types could provide new insight into the underlying biology, by guiding a priori GWAS analyses (e.g., brain enhancer elements for psychiatric disease) or facilitating post hoc interpretation.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Anthropometry
  • Autoimmune Diseases / genetics
  • Cardiovascular Diseases / genetics
  • Diabetes Mellitus / genetics
  • Gene Frequency
  • Genome-Wide Association Study*
  • Humans
  • Linkage Disequilibrium
  • Mental Disorders / genetics
  • Mutation, Missense
  • Neoplasms / genetics
  • Nervous System Diseases / genetics
  • Phenotype*
  • Polymorphism, Single Nucleotide*
  • Promoter Regions, Genetic
  • Quantitative Trait Loci