Experimental validation of data mined single nucleotide polymorphisms from several databases and consecutive dbSNP builds.

Rapid development in the annotation of human genetic variation has increased the numbers of single nucleotide polymorphisms (SNPs) in candidate genes by several orders of magnitude. The selection of both useful target SNPs for disease-gene association studies and SNPs associated with the treatment response is therefore an increasingly challenging task. We describe a workflow for selecting SNPs based on their putative function and frequency in candidate genes extracted from PubMed resources. The annotation of each SNP and its frequency in a Caucasian population was assessed in several databases. Approximately 4000 SNPs were identified from an initial 233 candidate genes. In a case study, we performed actual genotyping of 1030 of these SNPs in 213 genes and obtained 710 successfully genotyped SNPs. Using the flow-chart outlined here, only 87 SNPs were monomorphic (approximately 12%). This study reports the frequency of SNPs in a Caucasian population, selected in silico, using a candidate gene approach and validated by actually genotyping 193 individuals. The selected genotypes represent a valuable set of verified candidate SNPs for pharmacogenetic studies in Caucasian populations.