Human populations display enormous genetic variation, evident both at the phenotypic level and at the DNA level. These variations include both single "mutations" causative of profound disease and variations that contribute to susceptibility to particular "polygenic" diseases. To date, genetic diagnostics have been applied mainly in the context of isolated diagnoses for individuals or families for single gene disorders. Population screening, for example, for carrier status for recessive disorders such as cystic fibrosis, has not yet advanced beyond feasibility studies. Such studies require satisfaction of the criteria of Wilson and Jungner (1), in that there must be the knowledge and resources (counseling, treatment, and so on) to confer benefit, and that there must be a viable and affordable means to undertake screening. For research into the genetic epidemiology of common disorders, such as cardiovascular disease and cancers, some of the feasibility shortfalls correspond with those for population screening for single gene mutations. First, tests for genetic variation are still very expensive, both in staff time and in reagents. Second, the combination of informed consent, collection of appropriate clinical detail, and collection of samples is expensive on a population scale. For these reasons, many associations, such as those between apolipoprotein E (APOE) genotype, hypercholesterolaemia and premature vascular disease (2), APOE genotype and late-onset Alzheimer disease (3), and angiotensinogen-converting enzyme (ACE) genotype and cardiovascular disease (4), have mainly been researched on groups of a few hundred individuals.