Table V.
Ratio of frequencies (Yin:Yang) |
Pattern of Missing data 1 |
Simulated R1 (Yin haplotype) |
Power (1df Yin) |
Power (2df Yin+Yang) |
Empirical coverage of nominal 95% confidence region (2 risk parameters) |
Geometric mean of Estimated R1 of risk haplotype (95% CI)2 |
Error Rate3 |
|
---|---|---|---|---|---|---|---|---|
Yin | Yang | |||||||
1:1 (0.141:0.141) |
None | 1.414 | 0.730 | 0.631 | 0.954 | 1.42(1.41,1.43) | 1.00(1.00,1.01) | 0.040 |
0.707 | 0.617 | 0.506 | 0.950 | 0.71(0.70,0.71) | 1.00(0.99,1.00) | 0.043 | ||
20% g | 1.414 | 0.686 | 0.585 | 0.956 | 1.42(1.42,1.43) | 1.00(1.00,1.01) | 0.039 | |
0.707 | 0.574 | 0.467 | 0.948 | 0.71(0.70,0.71) | 1.00(0.99,1.00) | 0.042 | ||
1:3 (0.071:0.212) |
None | 1.414 | 0.479 | 0.373 | 0.949 | 1.42(1.41,1.43) | 1.00(1.00,1.01) | 0.036 |
0.707 | 0.385 | 0.297 | 0.949 | 0.70(0.70,0.71) | 1.00(1.00,1.00) | 0.033 | ||
20% g | 1.414 | 0.430 | 0.337 | 0.950 | 1.42(1.41,1.43) | 1.00(1.00,1.01) | 0.039 | |
0.707 | 0.344 | 0.261 | 0.950 | 0.70(0.70,0.71) | 1.00(1.00,1.00) | 0.037 |
None: no missing; 20% g: 20% of SNP genotypes are missing randomly.
The 95% CI is based on the empirical standard error calculated using the estimates from 5000 independent simulations.
Error rate incurred by designating the haplotype whose log relative risk deviated most from 0 as the risk-relevant one.