Multi-organ metabolome biological age implicates cardiometabolic conditions and mortality risk

Biological aging clocks across organs and omics data, including clinical phenotypes, neuroimaging, proteomics, and epigenetics, have proven instrumental in advancing our understanding of human aging and disease. Here, we expand this aging clock framework to plasma metabolomics by developing 5 organ-specific metabolome-based biological age gaps (MetBAGs) using 107 plasma non-derived metabolites from 274,247 UK Biobank participants. Our multi-organ MetBAGs were trained using Lasso regression and neural networks, achieving a mean absolute error of approximately 6 years (0.25<r<0.42) and comparable with literature for non-organ specific MetBAGs. Critically, including composite metabolites (e.g., sums or ratios of the original metabolites) reduced model generalizability to independent test data due to the high collinearity (or multicollinearity). Genome-wide associations identified 405 MetBAG-locus pairs (P<5x10{square}{square}/5). Using SBayesS, we estimated the SNP-based heritability (0.09<[Formula]<0.18), negative selection signatures (-0.93<S<-0.76), and polygenicity (0.001<Pi<0.003) for the 5 MetBAGs. Genetic correlation and Mendelian randomization analyses revealed potential causal links between the 5 MetBAGs and cardiometabolic conditions, such as metabolic disorders with the endocrine MetBAG and hypertension with the immune MetBAG. Integrating multi-organ and multi-omics features improved disease category and mortality predictions. The 5 MetBAGs expand upon existing biological aging clocks, providing an enriched framework for studying aging and disease across multiple biological scales.