Table 6.
Powera at a given threshold (180 families): proportion of times out of 1,000 simulations that the reported p value was more significant than the listed significance level (in percentages)
| RFGLS-UN | RFGLS-VC | FGLSb | GLS | OLSb | GWAFb | GEEb | QTDT | |
|---|---|---|---|---|---|---|---|---|
| α = 5 × 10−2 | ||||||||
| HomoG | 47.2 | 49.2 | 47.4 | 50.4 | 42.4 | 50.0 | 50.3 | 16.9 |
| HomoGE | 49.7 | 51.5 | 49.9 | 51.8 | 44.3 | 50.1 | 51.2 | 17.7 |
| HetGE | 44.6 | 44.8 | 44.8 | 47.1 | 37.6 | 45.5 | 44.6 | 15.0 |
| α = 5 × 10−3 | ||||||||
| HomoG | 16.3 | 19.4 | 16.3 | 19.7 | 13.6 | 19.7 | 18.2 | 2.7 |
| HomoGE | 19.1 | 21.6 | 19.7 | 21.3 | 15.3 | 19.5 | 20.2 | 4.2 |
| HetGE | 14.6 | 15.4 | 14.6 | 16.1 | 11.3 | 15.2 | 15.5 | 2.0 |
a
Power of detecting a single causal SNP which accounts for 0.6% of the variance.
b
The power of the FGLS, OLS, GWAF and GEE methods were calculated after the inflated type I error was corrected.