Figure 1.

Power (omnibus test) for a study of families, for two scenarios: R₁, R₂, S₁, and S₂, respectively, of 2, 4, 1.5, and 2 (upper curve) and of 1, 3, 2, and 2 (lower curve). (Calculations assumed a stratified population in which half have a baseline risk of 0.01 and an allele prevalence of 0.05 and the other half have a baseline risk of 0.03 and an allele prevalence of 0.10.) The omnibus null hypothesis is that the allele under study is not linked or not associated with a risk-conferring gene and that there are no maternally mediated genetic effects.