The ascertainment sampling problem and estimation of genetic parameters when parental haplotypes are known

Comparisons of bias and standard errors of genetic parameter estimates arising from various estimation procedures are made by using contrived genetic data in which parental haplotypes are known in addition to the haplotypes of affected sibs. These comparisons show that knowledge of parental haplotypes can reduce standard errors of genetic parameter estimates by between 10% and 45% as compared to estimates when parental information is not available. The added information does not appear to change the standard errors of ascertainment–assumption–free estimates relative to those from classical methods. This implies that the decision to accept possible bias arising from a classical estimation method or to accept slightly increased standard errors under an ascertainment–assumption–free approach is comparable to that involved when parental haplotypes are unknown.