A Cross-platform Approach Identifies Genetic Regulators of Human Metabolism and Health

The Challenge: Limitations of a single metabolomic platform

Blood metabolite levels are highly heritable, and a growing number of studies have begun to characterize the genetic landscape of metabolite fluctuations. However, previous reports examining genome-metabolite associations are often limited in size and scope due to the utilization of a single metabolomic platform. By combining results across various large-cohort studies utilizing different metabolomic platforms, optimal statistical power provides improved understanding of gene-metabolite associations, human physiology, and disease.

Metabolon Insight: A more robust genome-metabolite data set

To build a cross-platform metabolomic dataset, researchers utilized metabolites measured by several metabolomic platforms, including Metabolon’s Global Discovery Panel. This provided a set of platform-specific metabolites, enabling genome-wide meta-analyses on 174 metabolites from seven biochemical classes. With increased power and scope, researchers mapped gene-metabolite associations across platforms and included genetic architecture, determining putative causal genes for metabolite fluctuations and links to disease.

The Solution: Integration across metabolomic platforms

Among the metabolites analyzed, cross-platform metabolomics revealed 499 gene-metabolite associations from 144 loci. To determine the statistical viability of cross-platform integration, researchers investigated the effects of cohort, measurement platform, metabolite class, and association strength. The variabilities between platforms were largely due to the overall strength of the signal, but the directionality of associations was consistent, demonstrating that pooling measurements across metabolomic platforms is not only feasible but also enables more powerful analyses. Notably, this approach revealed that the genetic effects on metabolite levels are more than threefold compared to what is typically seen in common genetic variants.

Since many of the metabolites targeted in this study are associated with T2D, the researchers sought to understand whether this cross-platform approach could provide novel insights into T2D pathophysiology. Interestingly, results revealed variants in a receptor called GLP2R, and while common variants are linked to increased T2D risk, one particular variant (rs17681684, previously considered benign) was not only associated with T2D phenotypes but also with elevated plasma citrulline. These findings were extended through in vitro studies, where impaired β-arrestin-specific signaling due to this GLP2R variant could be a potential mechanism underlying T2D pathophysiology.

This study further revealed a strong link between serine levels and the development of macular telangiectasia type 2. Risk scores generated by gene-metabolite associations revealed that a one standard deviation increase in serine levels is associated with a 95% lower risk of developing this rare degenerative retinal disease. These results suggest that blood metabolite levels can be utilized to treat these types of diseases via supplementation or pharmacological manipulations.

The Outcome: New insights in type-2 diabetes and retinal disease

With the advent of mGWAS, this report emphasizes the value of integrating metabolome data across various measurement platforms. With bolstered statistical power, these findings provide a deeper understanding of gene-metabolite interactions that underlie human physiology and demonstrate the translational utility of gene-metabolite associations toward clinical applications.