Genetic epidemiology of single-nucleotide polymorphisms

On the causal hypothesis, most genetic determinants of disease are single-n ucleotide polymorphisms (SNPs) that are likely to be selected as markers fo r positional cloning. On the proximity hypothesis, most disease determinant s will not be included among markers but may be detected through linkage di sequilibrium with other SNPs, In that event, allelic association among SNPs is an essential factor in positional cloning. Recent simulation based on m onotonic population expansion suggests that,useful association does not usu ally extend beyond 3 kb. This is contradicted by significant disequilibrium at much greater distances, with corresponding reduction in the number of S NPs required for a cost-effective genome scan. A plausible explanation is t hat cyclical expansions follow population bottlenecks that establish new di sequilibria. Data on more than 1,000 locus pairs indicate that most disequi libria trace to the Neolithic, with no apparent difference between haplotyp es that are random or selected through a major disease gene. Short duration may be characteristic of alleles contributing to disease susceptibility an d haplotypes characteristic of particular ethnic groups. Alleles that are h ighly polymorphic in all ethnic groups may be older, neutral, or advantageo us, in weak disequilibrium with nearby markers, and therefore less useful f or positional cloning of disease genes. Significant disequilibrium at large distance makes the number of suitably chosen SNPs required for genome scre ening as small as 30,000, or 1 per 100 kb, with greater density (including less common SNPs) reserved for candidate regions.