The study, spanning EstBB and UKBB_EUR with nearly 600k participants, identifies 86,886 locus–trait associations across 8,260 independent lead variants, revealing substantial novelty and strong replication of many prior signals.

Genetic correlations among metabolic traits are generally modest, though strong within lipid traits and branched-chain amino acids; Mendelian randomization suggests many metabolic traits influence coronary artery disease and type 2 diabetes, while pleiotropy and complex causal relationships complicate interpretation.

(No additional key point provided in source data for this group.)

The analysis extends to low-frequency (0.1%–1%) and rare variants, identifying credible sets and fine-mapped variants with a higher fraction of coding or splice-impact predictions, highlighting regulators such as PCSK9, ABCA1, ANGPTL4, PAH, PDK3, and HAL.

The study provides a broad introductory overview of mapping genetic influences on 249 metabolic traits at scale across Estonian Biobank and UK Biobank, examining biology, correlations, and implications for disease across multiple ancestry groups.

There is convergence on the branched-chain amino acid catabolism pathway, with both common and rare variants in genes like BCAT1/2, BCKDHA/BCKDHB, DBT, DLD, BCKDK, and PPM1K showing associations, including novel rare missense and splice-site variants aligned with functional predictions.

Heritability estimates for the traits range from about 3% to 19.5% (median near 10%), with greater heritability linked to more discovered loci and most associations replicating in overlapping datasets, alongside abundant novel findings across traits.