Table 2.
Summary of the nominal type I error rate α and power of QTL detection 1−β ∗ of four statistical models combined with two cofactor selection methods (C1, C2) at different α ∗ levels in a scenario with 50 QTLs, heritability h 2= 0.8, and 40 backcross nested association mapping (BC-NAM) subpopulations which were randomly selected from a total of 50 BC-NAM subpopulations, where α is the mean nominal type I error rate across the performed 25 simulation runs, α ∗ is the empirical type I error rate, S1 and S2 refer to single marker model 1 and 2, J1 and J2 refer to joint composite interval mapping model 1 and 2. For details see ‘Methods’
| α ∗ | S1 | S2 | J1C1 | J2C1 | J1C2 | J2C2 | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| α | 1- β ∗ | α | 1- β ∗ | α | 1- β ∗ | α | 1- β ∗ | α | 1- β ∗ | α | 1- β ∗ | |||||||
| 0.00001 | 9.71×10−27 | 0.049 | 4.51×10−28 | 0.063 | 1.09×10−14 | 0.099 | 4.00×10−17 | 0.099 | 6.40×10−18 | 0.090 | 1.41×10−23 | 0.100 | ||||||
| 0.0001 | 9.68×10−26 | 0.052 | 5.69×10−28 | 0.064 | 5.01×10−14 | 0.102 | 3.99×10−16 | 0.105 | 5.35×10−17 | 0.096 | 1.41×10−22 | 0.101 | ||||||
| 0.001 | 6.06×10−19 | 0.100 | 1.54×10−21 | 0.113 | 5.05×10−11 | 0.188 | 2.86×10−13 | 0.190 | 3.87×10−12 | 0.177 | 3.31×10−16 | 0.184 | ||||||
| 0.01 | 6.41×10−11 | 0.238 | 2.50×10−12 | 0.248 | 5.17×10−5 | 0.398 | 1.68×10−5 | 0.433 | 3.23×10−5 | 0.391 | 5.25×10−6 | 0.417 | ||||||
| 0.05 | 1.33×10−6 | 0.405 | 3.84×10−7 | 0.429 | 1.00×10−2 | 0.600 | 8.13×10−3 | 0.660 | 1.20×10−2 | 0.620 | 8.79×10−3 | 0.695 | ||||||
| 0.1 | 8.12×10−5 | 0.506 | 3.53×10−5 | 0.544 | 4.22×10−2 | 0.696 | 3.86×10−2 | 0.747 | 4.94×10−2 | 0.706 | 4.51×10−2 | 0.768 | ||||||
| 0.5 | 1.34×10−1 | 0.830 | 1.17×10−1 | 0.844 | 4.42×10−1 | 0.908 | 4.37×10−1 | 0.920 | 4.57×10−1 | 0.896 | 4.54×10−1 | 0.917 | ||||||