Resequencing of 200 human exomes identifies an excess of low-frequency non-synonymous coding variants

Targeted capture combined with massively parallel exome sequencingis a promising approach to identify genetic variants implicated in human traits. We report exome sequencingof 200 individuals from Denmark with targeted capture of 18,654 coding genes and sequence coverage of each individual exomeat an average depth of 12-fold. On average, about 95% of the target regions were covered by at least one read. We identified 121,870 SNPs in the sample population, including 53,081 coding SNPs (cSNPs). Using a statistical method for SNP calling and an estimation of allelic frequenciesbased on our population data, we derived the allele frequencyspectrum of cSNPs with a minor allele frequencygreater than 0.02. We identified a 1.8-fold excess of deleterious, non-syonomyous cSNPs over synonymous cSNPs in the low-frequency range ( minor allele frequenciesbetween 2% and 5%). This excess was more pronounced for X-linked SNPs, suggesting that deleterious substitutions are primarily recessive.