Table 2. Actual type I error rates (%) of FITR.
T | L | Δta | Model 1 (R = 5)b | Model 2 (R = 2) | Model 3 (R = 20) | Model 4 (R = 1) | Model 5 (R = 10) |
---|---|---|---|---|---|---|---|
10 | 2 | 5 | 4.99 | 5.02 | 4.95 | 5.12 | 5.00 |
100 | 20 | 5 | 5.09 | 4.95 | 5.06 | 5.05 | 5.00 |
1000 | 200 | 5 | 5.07 | 4.95 | 5.01 | 5.02 | 5.24 |
10 | 5 | 2 | 4.98 | 4.94 | 5.06 | 4.95 | 4.94 |
100 | 5 | 20 | 4.98 | 5.17 | 5.20 | 4.90 | 5.02 |
1000 | 5 | 200 | 4.87 | 5.02 | 5.06 | 4.89 | 5.02 |
Values indicate the actual type I error rates obtained by 100,000 simulations under a nominal significance level of 5%. The initial frequencies for all R +1 loci, xh,0, are assumed to be 0.5. FITR, frequency increment test with reference loci.
.
The numbers of reference loci, R, are randomly assigned to each demographic model.