Functional genomics analysis of big data identifies novel PPARy target SNPs showing association with cardio metabolic outcomes

Authors: RICHARDSON, KRIS - Jean Mayer Human Nutrition Research Center On Aging At Tufts University; SCHNITZLER, GAVIN - Tufts - New England Medical Center; Lai, Chao Qiang; ORDOVAS, JOSE - Jean Mayer Human Nutrition Research Center On Aging At Tufts University

Submitted to: Circulation: Cardiovascular Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/22/2015
Publication Date: 12/1/2015
Citation: Richardson, K., Schnitzler, G.R., Lai, C., Ordovas, J.M. 2015. Functional genomics analysis of big data identifies novel PPARy target SNPs showing association with cardio metabolic outcomes. Circulation: Cardiovascular Genetics. 8(6):842-851.

Interpretive Summary: Cardiovascular disease and type-2-diabetes represent overlapping diseases and despite hundreds of GWAS (genetic screens for common mutations) and meta-analytic studies by Consortia, a large portion of the variation attributable to genetics remains unexplained. An important player in their etiology is Peroxisome Proliferator-activated Receptor gamma (PPARy) that is involved in lipid and glucose metabolism and maintenance of metabolic homeostasis. Recent seminal papers have demonstrated the importance of transcriptionally active intergenic regions of the genome in the regulation of metabolic processes. Along with these conclusions were published a wealth of raw ChIP-Seq data for dozens of transcription factors, thus allowing researchers to fine map the corresponding functional genomic elements. Here we demonstrate how an a priori Functional Genomics approach using Big Data can provide insight into the molecular mechanisms underlying a variants association with a disease outcome. Specifically, we have integrated data on the binding of PPARg to its DNA targets, cardio-metabolic GWAS data and eQTL data derived from human adipose. With these datasets we identified common human genetic variants falling within PPARg binding sites that associate with one or more cardio-metabolic outcome in addition to allele specific gene-expression changes. This information may be further utilized to select high priority common mutation candidates for functional studies and may provide predictive, diagnostic or therapeutic information to health care professionals.

Technical Abstract: Background - Cardiovascular disease and type-2-diabetes represent overlapping diseases where a large portion of the variation attributable to genetics remains unexplained. An important player in their etiology is Peroxisome Proliferator-activated Receptor gamma (PPARy) that is involved in lipid and glucose metabolism and maintenance of metabolic homeostasis. We utilized a functional genomics methodology to interrogate human ChIP-Seq, GWAS and eQTL data to inform selection of candidate functional SNPs falling in PPARy motifs.Methods and Results We derived 27328 ChIP-Seq peaks for PPARy in human adipocytes through meta analysis of three datasets. The PPARy consensus motif showed greatest enrichment and mapped to 8637 peaks. We identified 146 SNPs in these motifs. This number was significantly less than would be expected by chance, and INSIGHT analysis indicated these motifs are under weak negative selection. A screen of these SNPs against GWAS for cardio-metabolic traits revealed significant enrichment with 16 SNPs. A screen against the MuTHER eQTL data revealed 8 of these were significantly associated with altered gene expression in human adipose, more than would be expected by chance. Several SNPs fall close, or are linked by eQTL, to lipid-metabolism loci including CYP26A1.Conclusions We demonstrated the utility of functional genomics to identify SNPs of potential function. Specifically, that SNPs within PPARy motifs that bind PPARy in adipocytes are significantly associated with cardio-metabolic disease and with the regulation of transcription in adipose. This method may be used to uncover functional SNPs that do not reach significance thresholds in the agnostic GWAS approach.