Simple simulated datasets were created for checking the effective sample size method as applied to sibpair data. For each replicate, genotypes of 500 “case” samples consisting of 250 sibpairs and 500 “control” samples were simulated. The genotype in control group was sampled from the genotype distribution of Pcontrol(G) = (p2, 2pq, q2) for genotypes AA, AB, BB. Those in the case group is sampled by the model: (14)Pcase(G)∝ea+b·f(G)1+ea+b·f(G)Pcontrol(G) where f(G) represents the disease models, a is the baseline log-odds, and b is the log-odds ratio. The dominant model (D) is equivalent to f(G)=(1,1,0) for genotype AA, AB, BB; recessive model (R) corresponds to f(G)=(1,0,0); and additive model (A) corresponds to f(G) = (1, 0.5, 0). For the null distribution to be used to check the type I error, b = 0, i.e, genotype has no effect on the disease status, and Pcase = Pcontrol. For the alternative distribution to be used to check the power, a is chosen at −4 and b is chosen between 0 and 0.5.
