The usefulness of different density SNP maps for disease association studies of common variants

Large-scale discovery and validation of single-nucleotide polymorphisms (SNPs) facilitates indirect association mapping. It has recently been estimated that, in Europeans, 77% of all SNPs with frequency of 10% or more could be ascertained through linkage disequilibrium (LD) by genotyping variants in the database dbSNP. Using a sampling approach from 73 genes with near complete SNP maps, we show here the usefulness of SNP maps at different densities and the large variability of SNP coverage in different genomic regions. While even sparse SNP maps are of some value to genetic mapping, in order to undertake disease association studies providing at least 80% of SNPs in 90% of genes, much denser maps need to be constructed, at more than one SNP per kb in some regions.