Single nucleotide polymorphisms (SNPs) are highly abundant in the genome and especially useful in the search for disease susceptibility genes via population-based association or linkage studies. Therefore, there is a strong need for high throughput and reliable methodologies to assess the SNP genotypes. Despite an unambiguous result of an SNP analysis, with the use of a commercial kit based on primer extension, subsequent sequencing analysis revealed that a proportion of the genotypes was not correctly assessed. The problem we have encountered may originate from specific structures in the genomic DNA sequence, rather than being a methodological problem.