Table IV.

Performance of the two-candidate-haplotype log-linear model enforcing HWE. Genotypes were simulated mimicking RFC1 haplotypes under HWE with two haplotypes conferring risk: haplotype 1, R₁ =1.2 and R₂=R₁²; haplotype 2, T₁= 1.35 and T₂=T₁². Each scenario reflects 5000 simulations.

N	Pattern of Missing data 1	Simulated haplotype 1 frequency	Simulated haplotype 2 frequency	Power of test including only one haplotype as candidate2		Power of test including both haplotypes as candidates	Empirical coverage of nominal 95% confidence region	Geometric mean of estimated R1, T1 (both haplotypes included as candidate haplotypes in the model, 95% CI)3
				haplotype 1	haplotype 2			haplotype 1	haplotype 2
400	None	0.141	0.141	0.152	0.500	0.537	0.956	1.20(1.20,1.21)	1.35(1.35,1.36)
	20% g			0.127	0.492	0.514	0.956	1.20(1.20,1.21)	1.35(1.35,1.36)
	None	0.0707	0.212	0.093	0.671	0.632	0.953	1.21(1.20,1.21)	1.35(1.35,1.35)
	20% g			0.082	0.663	0.620	0.951	1.21(1.20,1.21)	1.35(1.35,1.35)
	None	0.212	0.0707	0.235	0.283	0.420	0.947	1.20(1.19,1.20)	1.35(1.34,1.36)
	20% g			0.212	0.276	0.399	0.952	1.20(1.20,1.20)	1.35(1.34,1.36)
1000	None	0.141	0.141	0.292	0.868	0.915	0.948	1.20(1.20,1.20)	1.35(1.35,1.35)
	20% g			0.243	0.865	0.904	0.950	1.20(1.20,1.20)	1.35(1.35,1.35)
	None	0.0707	0.212	0.138	0.965	0.960	0.950	1.20(1.20,1.21)	1.35(1.35,1.35)
	20% g			0.105	0.962	0.955	0.949	1.20(1.20,1.21)	1.35(1.35,1.35)
	None	0.212	0.0707	0.498	0.598	0.833	0.942	1.20(1.20,1.20)	1.35(1.35,1.36)
	20% g			0.446	0.588	0.803	0.943	1.20(1.20,1.20)	1.35(1.35,1.36)

¹None: no missing; 20% g: 20% of SNP genotypes are missing randomly.

²Only one risk haplotype is included as the candidate haplotype and the other is aggregated with the non-risk haplotypes.

³The 95% CI is based on the empirical standard error calculated using the estimates from 5000 independent simulations.