Nuclear magnetic resonance spectroscopy has been exploited to study the metabolic characteristics (phenotype) of genetic disorders by taking advantage of some unique characteristics of the technique. The first application, metabolic profiling for diagnosis and therapeutic monitoring in vitro, demonstrates the exceptional diversity of metabolites detected by NMR, and has resulted in new interest in significant metabolites largely ignored previously because other techniques do not detect them, e.g. betaine and creatine. Moreover, previously 'unknown' genetic disorders have been detected and characterised The same NMR technique can be effectively exploited for metabolic profiling of mutation models in yeast and mice, leading to a prominent role in the development of large scale metabolomic profiling to link genomic information with phenotype. The second application, magnetic resonance spectroscopy (MRS), exploits the unique possibility of studying human metabolism in vivo, which permits intracellular rather than extracellular metabolic profiling. When it is possible to detect the precise diagnostic metabolites in vivo, investigators have been able to link clinical status with cellular biochemistry, sometimes questioning the clinical value of extracellular (plasma) metabolite measurements. Thus, claims have been made that brain phenylalanine concentrations match more closely the clinical status of patients with phenylketonuria. These studies in vivo have also led to new diagnoses e.g. the disorders of creatine synthesis and transport, highlighting a new category of brain syndromes. Future applications of NMR are cautiously considered as they are critically dependent on continued improvement in resolution and sensitivity in turn generated by developments in magnet design and higher fields.