True pedigree errors more frequent than apparent errors for single nucleotide polymorphisms

Single nucleotide polymorphisms (SNPs) are currently being developed for us e in disequilibrium analyses. These SNPs consist of two alleles with varyin g degrees of polymorphism. A natural design for use with SNPs is the 'haplo type relative risk' sampling design in which a father, mother, and child ar e typed at an SNP locus. Given such a trio of genotypes, we ask: what is th e probability that a pedigree error (a change from one allele to the other) at an SNP locus will be detected using only Mender's laws as a check? We c alculate the probability of detecting such errors for a hypothetical SNP lo cus with varying degrees of polymorphism and for various true error rates. For the sets of allele frequencies considered, we find that the detection r ates range between 25 and 30%, the detection rate being lowest when the two alleles have equal frequencies and the highest when one allele has a frequ ency of 10%. Based on this detection rate, we determine that the true error rate is roughly 3.3-4 times that of the apparent error rate at an SNP locu s. The greatest discrepancy between true and apparent error rates occurs wh en allele frequencies are equal.