Figure 4.

Power comparison of case-control design and one sibling per ASP–control design, under multilocus disease models, when the effect size of each disease locus is fixed and the number of disease loci increases. Results are based on 2,000 replicate data sets. The disease prevalence K=5%. The disease is influenced by L (2⩽L⩽10) unlinked disease loci, each with a predisposing-allele frequency of 0.1. The SNP, with a minor-allele frequency of 0.1, is completely linked to the first disease locus, and r² between the two loci is 0.5. All disease loci follow a dominant, an additive, or a recessive model, with locus-specific λ_s at each disease locus fixed at 1.02. Power is assessed at the 1% level. The solid line is for design with 500 cases (one sibling per ASP) and 500 controls. The dashed line is for design with 500 cases and 500 controls.