. Author manuscript; available in PMC: 2014 Nov 20.

Published in final edited form as: Genet Epidemiol. 2011 Sep 15;35(8):790–799. doi: 10.1002/gepi.20628

TABLE I.

Power of extreme two-stage and traditional two-stage designs with different proportion of subjects (π_Sample) allocated in the stage1^a

	β = 0.71^b		β = 0.50^c		β = 0.35^d
π_Sample	Extreme two-stage	Traditional two-stage	Extreme two-stage	Traditional two-stage	Extreme two-stage	Traditional two-stage
0.05	0.52	0.03	0.50	0.03	0.48	0.03
0.1	0.78	0.11	0.77	0.11	0.74	0.11
0.2	0.89	0.36	0.88	0.36	0.88	0.36
0.4	0.90	0.76	0.90	0.76	0.90	0.76
0.6	0.90	0.88	0.90	0.88	0.90	0.88

The proportion of markers selected to follow up in stage 2 (π_marker) is set to be 0.001. The MAF of the causal SNP is set to be 0.005. Genome-wide significance level is set to be 5.0 × 10⁻⁸. In the extreme two-stage design, we assume individuals from π_sample/2 proportion on both tails of the phenotype are located in the first stage; in the traditional two-stage design, π_sample proportion of individuals is randomly located in the first stage.

Corresponding to a variance contribution of 0.5%, total sample size is set to be 9,100 to ensure a maximum power of 0.90.

Corresponding to a variance contribution of 0.25%, total sample size is set to be 18,177 to ensure a maximum power of 0.90.

Corresponding to a variance contribution of 0.125%, total sample size is set to be 36,331 to ensure a maximum power of 0.90.