Estimating genotyping error rates from parent-offspring dyads

A common approach when estimating the error rate of a DNA register is to genotype some of the individuals twice. This may be both expensive and time consuming. As an alternative, we present a new method for estimating genotyping errors based on parent-offspring dyads. The basic idea is that parent and offspring must share at least one allele per locus. Others have previously devised similar techniques, but depended on the assumption that at most one error may occur per dyad. In this paper we examine the bias caused by this simplification. Further, we apply our method on a data set from the Norwegian minke whale DNA register, and find that the error rates are in the range of 0-0.0418, which is comparable to those in the published literature.