Massive GWAS maps 249 circulating metabolites to thousands of genetic signals in 619,372 individuals
A large GWAS meta-analysis of 249 circulating metabolic traits in the Estonian Biobank and UK Biobank (up to 619,372 individuals) identifies 88,127 locus–trait associations across 8,398 loci. Using fine mapping, phenome‑wide colocalization, and cis‑Mendelian randomization, the study highlights the value of low‑frequency variants (MAF 0.1–1%) in explaining heritability and points to causal links between metabolic traits and diseases such as CAD and T2D. Notably, 19.4% of confidently fine‑mapped variants are low‑frequency missense or splice variants, enriching interpretability. The work also uncovers three lactate‑related loci (GP6, GRK5, ZFPM2) where higher plasma lactate associates with pulmonary embolism risk, potentially reflecting platelet activation rather than a direct causal effect. Overall, integrating low‑frequency variation improves mapping precision and biological insight into metabolism and disease.