Widespread purifying selection at polymorphic sites in human protein-coding loci.

Estimation of gene diversity (heterozygosity) at 1442 single-nucleotide polymorphism (SNP) loci in an ethnically diverse sample of humans revealed consistently reduced gene diversities at SNP loci causing amino acid changes, particularly those causing amino acid changes predicted to be disruptive to protein structure. The reduction of gene diversity at these SNP loci, in comparison to SNPs in the same genes not affecting protein structure, is evidence that negative natural selection (purifying selection) has reduced the population frequencies of deleterious SNP alleles. This, in turn, suggests that slightly deleterious mutations are widespread in the human population and that estimation of gene diversity even in a sample of modest size can help guide the search for disease-associated genes.