Inborn errors of metabolism (IEMs) are a group of disorders characterised by the toxic accumulation or deficiency of circulating molecules (‘metabolites’) caused by rare genetic mutations. Previous studies have identified select examples where common variants at genes known to cause rare Mendelian diseases, including IEMs (e.g. LPL, DBH, PPM1K), are linked to phenotypic consequences in the general population that also occur in patients with the corresponding rare disease. Advances in genetic and metabolic profiling at an epidemiological scale now provide an opportunity to systematically identify such examples in the population and characterise their downstream effects on health. To assess the value of untargeted metabolomic profiling for the study of common complex diseases, I identified candidate mediators of the association between weight gain and future type 2 diabetes risk based on untargeted, large-scale metabolomic profiling of a large prospective cohort. Integration of metabolomics, genetic profiling and comprehensive longitudinal follow up for a range of diseases together with the application of Bayesian and genetic epidemiological methods enabled the identification of 20 candidate mediators. These reflected genetic susceptibility to adiposity and insulin resistance and explained most of the increased T2D risk associated with weight gain. To systematically characterise the phenotypic effects of variation at IEM-causing genes, I identified sentinel variants at these genes associated with plasma metabolites affected in the corresponding IEM across the genome. Of the 202 ‘IEM familiar’ variants (IFVs) detected, 187 at 89 loci were not previously reported as pathogenic for the corresponding IEM in ClinVar and 51 of these were associated with extreme metabolite levels (<2.5th or >97.5th percentile) or had non-additive effects on metabolite levels. Phenome-wide assessment identified 1,553 IFV-phenotype associations at 108 loci. Of the detected associations, 703 at 54 loci were of particular interest as the phenotype related to a symptom of the corresponding IEM. At 24 of these 54 loci, genetic colocalisation detected shared genetic signals for IEM-related metabolites and phenotypes. For example, in line with norepinephrine deficiency causing dizziness on standing in severe cases of rare orthostatic hypotension (OMIM #223360), I identified a genetic signal at the dopamine beta hydroxylase (DBH) locus associated with decreased levels of the downstream catecholamine vanillylmandelate in the general population (IFV EAF=0.074). This signal was shared with that for lower risk of hypertension (based on 462,933 participants in UK Biobank) and other blood pressure-related phenotypes with high posterior probability of colocalisation (PPcolocalisation=0.94, with >99% of the probability explained by the IFV). This work uses untargeted metabolomic profiling to identify underlying disease mechanisms and demonstrate the proof-of-principle that common variants can have similar health consequences to those caused by rare mutations at the same IEM gene.