Discovery and Fine-Mapping of Glycaemic and Obesity-Related Trait Loci Using High-Density Imputation

Abstract

Reference panels from the 1000 Genomes (1000G) Project Consortium provide near complete coverage of common and low-frequency genetic variation with minor allele frequency ≥0.5% across European ancestry populations. Within the European Network for Genetic and Genomic Epidemiology (ENGAGE) Consortium, we have undertaken the first large-scale meta-analysis of genome-wide association studies (GWAS), supplemented by 1000G imputation, for four quantitative glycaemic and obesity-related traits, in up to 87,048 individuals of European ancestry. We identified two loci for body mass index (BMI) at genome-wide significance, and two for fasting glucose (FG), none of which has been previously reported in larger meta-analysis efforts to combine GWAS of European ancestry. Through conditional analysis, we also detected multiple distinct signals of association mapping to established loci for waist-hip ratio adjusted for BMI (RSPO3) and FG (GCK and G6PC2). The index variant for one association signal at the G6PC2 locus is a low-frequency coding allele, H177Y, which has recently been demonstrated to have a functional role in glucose regulation. Fine-mapping analyses revealed that the non-coding variants most likely to drive association signals at established and novel loci were enriched for overlap with enhancer elements, which for FG mapped to promoter and transcription factor binding sites in pancreatic islets, in particular. Our study demonstrates that 1000G imputation and genetic fine-mapping of common and low-frequency variant association signals at GWAS loci, integrated with genomic annotation in relevant tissues, can provide insight into the functional and regulatory mechanisms through which their effects on glycaemic and obesity-related traits are mediated.

Fig 1. Regional plots of multiple distinct signals at WHR_adjBMI locus RSPO3 (A), FG loci G6PC2 (B) and GCK (C).

Regional plots for each locus are displayed from: the unconditional meta-analysis (left); the exact conditional meta-analysis for the primary signal after adjustment for the index variant for the secondary signal (middle); and the exact conditional meta-analysis for the secondary signal after adjustment for the index variant for the primary signal (right). The sample sizes vary due to the availability of the well imputed index SNPs of the primary and secondary signals. Directly genotyped or imputed SNPs are plotted with their association P values (on a -log₁₀ scale) as a function of genomic position (NCBI Build 37). Estimated recombination rates are plotted to reflect the local LD structure around the associated SNPs and their correlated proxies (according to a blue to red scale from r ² = 0 to 1, based on pairwise EUR r ² values from the 1000 Genomes June 2011 release). SNP annotations are as follows: circles, no annotation; downward triangles, nonsynonymous; squares, coding or 3′ UTR; asterisks, TFBScons (in a conserved region predicted to be a transcription factor binding site); squares with an X, MCS44 placental (in a region highly conserved in placental mammals).

Fig 2. Broad category functional annotation (A) and cell-type specific annotation (B) of credible set variants.

On the x-axis is each category of broad functional annotation (A) or cell-type specific annotation (B). The fraction of credible set variants that overlap with each category is shown on y-axis. The overlapping variants are further broken down into either variants that exist in both the 1000 Genomes and HapMap reference panel (green) or those that exist only in the 1000 Genomes reference panel (red). TFBS, transcription factor binding site; ncRNA, non-coding RNA; UTR, untranslated regions; GM12878, lymphoblastoid cell line from European ancestry female; hESC, H1 human embryonic stem cells; hASC(t1), human pre-adipocytes; hASC(t4), mature human adipocytes; HepG2, liver carcinoma cell-line; HMEC, human mammary epithelial cells; HSMM, human skeletal muscle myoblasts; HUVEC, human umbilical vein endothelial cells; K562, human myelogenous leukemia cell-line; NHEK, normal human epidermal keratinocytes; NHLF, normal human lung fibroblasts.