Case Study
Whole-genome sequencing identifies common-to-rare variants associated with human blood metabolites
Combining whole genome sequencing and metabolomic profiling of a subset of the Twins UK cohort enabled the identification of several heterozygous rare variants associated with abnormal metabolite levels in adulthood.
In a large collaborative effort led by senior researchers J. Craig Venter, Timothy Spector, and Amalio Telenti, researchers leveraged the thousands of individuals in the Twins UK cohort with deep whole genome sequencing data to identify several heterozygous rare variants associated with inborn errors of metabolism or other pediatric genetic conditions. This study, published in Nature Genetics, combines whole genome sequencing data with state-of-the-art metabolic profiling to describe for the first time the prevalence of abnormal metabolite levels in adults with heterozygous rare variants, extending on existing datasets mapping genetic loci to the human metabolome.
In a large collaborative effort led by senior researchers J. Craig Venter, Timothy Spector, and Amalio Telenti, researchers leveraged the thousands of individuals in the Twins UK cohort with deep whole genome sequencing data to identify several heterozygous rare variants associated with inborn errors of metabolism or other pediatric genetic conditions. This study, published in Nature Genetics, combines whole genome sequencing data with state-of-the-art metabolic profiling to describe for the first time the prevalence of abnormal metabolite levels in adults with heterozygous rare variants, extending on existing datasets mapping genetic loci to the human metabolome.
The Challenge: Missing links between genetic traits and blood metabolites
Research has provided evidence of the genetic heritability of several metabolites reflecting the diversity of genes that encode enzymes, transporters, and other proteins involved in metabolism. Quantitative and accurate measurement of metabolic traits through metabolomics has enabled the mapping of loci and causal genes—including some rare variants—that influence a wide variety of metabolites. These studies have been performed on relatively small cohorts, however, and therefore haven’t been able to capture the full genetic picture. By combining whole genome sequencing data and metabolome profiling in a large cohort of nearly 2000 individuals, researchers were able to identify individual variants influencing metabolite levels (metabolic quantitative trait loci; mQTLs) and connect those mQTLs to several genetic conditions.
The Metabolon Insight: A robust gene-metabolome dataset compiled on hundreds of genetically identical individuals
The serum samples of 1960 participants in the Twins UK study were collected across three visits over a period of 18 years and analyzed using Metabolon’s Global Discovery Panel. A total of 901 metabolites covering eight broad biochemical classes were identified. Leveraging the presence of hundreds of pairs of twins in the Twins UK cohort, the researchers determined the heritability of several metabolites and used GWAS analysis to identify several heterozygous rare variants associated with key metabolites.
The Metabolon Insight: A robust gene-metabolome dataset compiled on hundreds of genetically identical individuals
Of the 901 metabolites identified using Metabolon’s platform, 644 were at consistent levels across all three visits and exhibited measurable heritability from 10.5% (methionine sulfoxide) to as much as 93.2% (ethylmalonate). This heritability emphasized the utility of metabolite levels as phenotypes for genetic association analysis and therefore, the researchers next performed GWAS on 6.69 million common and 4.66 million rare variants for each of the The researchers identified 223 variants independently associated with one or more of 246 metabolites, including associations with variants mapped to ACADS and NAT8, corroborating previous studies. Of these, 90 variants across 60 genetic loci represented novel mQTLs. Furthermore, the structures of some novel metabolites were imputed by analyzing the genetic variants to which 58 of these novel metabolites were associated.
Next, focusing their efforts on 151 individuals in the cohort with metabolite levels considered “extreme outliers” compared the rest of the study population, the researchers identified 14 rare variants across 10 genes associated with these metabolites. Detection of five of these variants across unrelated individuals supported their causality to the observed metabolite levels. A deeper analysis of 93 individuals carrying the 14 rare variants in 10 genes revealed that all individuals were heterozygous carriers of the variants.
Five of the 10 genes (UMPS, DMGDH, ACADS, ETFDH, and CRAT) are associated with inborn errors of metabolism, including orotic aciduria, dimethylglycine-dehydrogenase deficiency, short-chain acyl-CoA dehydrogenase deficiency (SCADD), glutaric acidemia type II (GA-II), and carnitine acetyltransferase deficiency. Notably, SCADD and GA-II are included in current newborn screening in the U.S. The fact that variants associated with these conditions were associated with abnormal metabolite levels in individuals heterozygous for these variants is significant, considering that heterozygosity in these genes is typically assumed to be asymptomatic.
Rare variants were associated with several other metabolic pathways and diseases, including lactase deficiency leading to neonatal diarrhea, abnormal 1,5-anhdroglucitol associated with hyperglycemia, and dopamine reuptake in infantile parkinsonism dystonia. Notably, the variant associated with infantile parkinsonism dystonia (SLC6A3; heterozygous) was observed in a single individual with adult-onset Parkinson’s disease.
The Outcome: A comprehensive variant-metabolite association map
This study is the largest known effort combining GWAS analysis with comprehensive metabolite mapping of human serum samples. By leveraging twins, this study enabled the identification of several heterozygous rare variants associated with abnormal metabolite levels in adults and with known connections to pediatric genetic conditions, including inborn errors of metabolism. This study is a brilliant example of how the accuracy of metabolite measurements through untargeted mass spectrometry and the whole genome coverage (including noncoding regions) provided by GWAS can contribute to our understanding of the relationship between genes and metabolites in human health and disease.