NEW YORK – With the help of exome sequence data generated for the UK Biobank study, a team led by researchers at AstraZeneca has obtained a more comprehensive view of the protein-coding variants that influence blood metabolites and related clinical markers.
"By exploring the full allelic frequency spectrum of protein-coding variation in relation to a range of metabolic blood measurements, our study expands the understanding of the genetic architecture of these measures in the UKB," senior author Slavé Petrovski, head of AstraZeneca's Centre for Genomics Research, and his colleagues reported in the American Journal of Human Genetics on Monday.
Building on findings from past genome-wide association studies, researchers in the UK, US, Sweden, and Australia analyzed exome sequences for nearly 412,400 UKB participants of European, South Asian, African, or East Asian ancestry to search for rare protein-coding variants associated with 355 blood metabolite measurements. These included 30 routinely tested clinical blood biomarkers and 325 metabolites profiled with nuclear magnetic resonance (NMR) spectroscopy-based assays by Finland's Nightingale Health in about 100,000 UKB participants.
With this approach, the team flagged some 1,968 blood metabolite associations, along with 331 associations involving clinical biomarkers in the blood. These spanned 205 different genes and expanded the suite of rare coding variants linked to clinical biomarkers or metabolite measures.
"The availability of exome sequences in the UKB now allows for a more systematic exploration of protein-coding variants across the allele frequency spectrum (including ultra-rare to rare variants) to identify associations that may either underpin known GWAS loci or represent previously unreported loci regulating metabolic blood measurements," the authors explained.
In an email, Petrovski noted that the availability of protein-coding sequences provided a threefold boost in the team's ability to find significant associations compared to approaches that involved imputed genotypes for the UK Biobank cohort.
In contrast, he explained, "only a fraction of our statistically significant signals were identified in prior studies that were either well-powered but limited to the study of common variant contributions or studies that focused on rare variants but in considerably smaller sample sizes."
Along with potentially causal variants in genes with well-established ties to metabolic traits or conditions, the analyses highlighted previously unappreciated candidate genes and variants.
The team flagged rare SYT7 gene variants with ties to blood levels of the kidney function marker creatinine, for example, as well as rare, non-synonymous variants in genes such as PLIN1 and CREB3L3 linked to lipid metabolite levels.
Consistent with PLIN1's role in lipid-related conditions, the investigators reported that protein-truncating variants in PLIN1 appeared to coincide with potentially beneficial lipid profiles, including lower-than-usual triglyceride levels, enhanced high-density lipoprotein levels, and reduced atherosclerotic heart disease incidence in a phenome-wide association analysis of UKB participants.
On the other hand, increased representation of "extremely large" forms of risky very-low-density lipoprotein (VLDL) cholesterol was associated with protein-truncating variants in CREB3L3, a gene previously implicated in fatty liver disease.
"This finding supports the notion that the link between CREB3L3 and fatty liver disease may in fact be a manifestation of metabolic syndrome in general, since the overproduction of large VLDL particles is a known hallmark of metabolic syndrome," the authors wrote.
More broadly, they suggested that the findings of the study "will not only be an important resource to identify therapeutic targets but also provide insights into the metabolic biomarker profile for human genetic perturbations that might be of interest in preclinical development."