Population-Scale Variant Analysis Reveals Novel Regulatory Elements Associated with Metabolic Syndrome Risk
¹ Computational Genomics Lab, Institute for Precision Medicine, Stanford University, CA, USA
² Department of Bioinformatics, Karolinska Institute, Stockholm, Sweden
³ Division of Endocrinology, Massachusetts General Hospital, Boston, MA, USA
Abstract
Background
Metabolic syndrome (MetS), affecting approximately 25% of adults globally, is a cluster of conditions including central obesity, insulin resistance, dyslipidemia, and hypertension. While genome-wide association studies (GWAS) have identified numerous common variants associated with individual MetS components, the regulatory mechanisms linking non-coding variants to disease risk remain incompletely characterized.
Methods
We performed a multi-ancestry GWAS meta-analysis integrating whole-genome sequencing data from 487,000 individuals across six population cohorts (European, East Asian, African, South Asian, Hispanic/Latino, and Middle Eastern ancestry). Phenotypic harmonization included standardized diagnostic criteria for MetS components. To prioritize functional variants, we integrated chromatin accessibility (ATAC-seq), histone modification (ChIP-seq), and gene expression (RNA-seq) data from 18 metabolically relevant tissues obtained from GTEx v9 and ENCODE.
Computational variant-to-gene mapping employed Activity-by-Contact (ABC) modeling and fine-mapping with SuSiE to identify credible causal variants. Biological validation was conducted via CRISPR interference (CRISPRi) screens in human hepatocyte and adipocyte cell lines.
Results
Our analysis identified 247 genome-wide significant loci (P < 5×10⁻⁸) associated with composite MetS risk, including 89 novel signals not previously reported. Fine-mapping resolved 18 loci to single putative causal variants. Functional integration prioritized 412 regulatory variants predicted to modulate expression of 196 genes enriched for lipid metabolism, glucose homeostasis, and inflammatory response pathways.
We identified a cluster of variants in a 15 kb enhancer region on chromosome 11q23.3, 340 kb upstream of FADS1/FADS2 (fatty acid desaturase genes), exhibiting strong association with triglyceride levels (lead SNP rs174547, OR=1.23, P=3.2×10⁻⁴⁴) and MetS risk (OR=1.18, P=7.1×10⁻²⁸). This region demonstrated open chromatin and H3K27ac enrichment specifically in liver tissue. CRISPRi-mediated silencing of this enhancer in hepatocytes reduced FADS1 expression by 47% and altered cellular fatty acid composition.
Multi-ancestry analysis revealed substantial effect size heterogeneity, with 34 loci showing ancestry-specific associations (Phet < 0.001). For example, the APOE ε4 allele conferred higher MetS risk in European ancestry participants (OR=1.31) compared to East Asian ancestry (OR=1.09, Pinteraction=2.4×10⁻⁵).
Conclusions
This study substantially expands the catalog of MetS-associated genetic variants and demonstrates the power of integrating population-scale genomic data with functional genomics to elucidate regulatory mechanisms. Our findings nominate specific non-coding variants and their target genes as potential therapeutic entry points for metabolic disease intervention. The observed ancestry-specific effects underscore the importance of diverse cohorts in genetic studies and suggest that precision medicine approaches for MetS may require population-tailored risk models.
Presentation Preference