TABLE 2.
Proportion of 95% confidence intervals that do not contain trueNe
|
Ne = 20
|
Ne = 50
|
Ne = 100
|
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Conditions | T | L | U | Tot | N̂ȯe | L | U | Tot | N̂ȯe | L | U | Tot | N̂ȯe |
| 5 loci | |||||||||||||
| n = 20 | 1 | 0.0 | 2.1 | 2.1 | 8.9 | 0.0 | 1.0 | 1.0 | 37.8 | ||||
| 3 | 4.5 | 1.5 | 6.0 | 0.1 | 0.5 | 1.2 | 1.7 | 7.2 | |||||
| 5 | 3.9 | 1.0 | 4.9 | 0.0 | 4.5 | 1.1 | 5.6 | 0.7 | |||||
| 10 | 2.9 | 3.0 | 5.9 | 0.0 | 7.2 | 1.0 | 8.2 | 0.0 | |||||
| n = 60 | 1 | 1.6 | 4.2 | 5.8 | 0.1 | 0.0 | 1.3 | 1.3 | 6.2 | 0.0 | 1.2 | 1.2 | 26.8 |
| 3 | 1.9 | 4.6 | 6.5 | 0.0 | 3.8 | 2.1 | 5.9 | 0.1 | 2.7 | 1.4 | 4.2 | 1.6 | |
| 5 | 1.6 | 4.5 | 6.1 | 0.0 | 3.8 | 2.8 | 6.6 | 0.0 | 4.2 | 1.9 | 6.1 | 0.4 | |
| 10 | 1.1 | 4.8 | 5.9 | 0.0 | 3.3 | 3.3 | 6.6 | 0.0 | 4.1 | 2.2 | 6.3 | 0.0 | |
| 15 loci | |||||||||||||
| n = 20 | 1 | 6.9 | 0.6 | 7.5 | 1.4 | 0.0 | 0.3 | 0.3 | 29.6 | ||||
| 3 | 5.6 | 1.2 | 6.8 | 0.0 | 10.8 | 0.3 | 11.1 | 0.8 | |||||
| 5 | 4.6 | 1.1 | 5.7 | 0.0 | 9.4 | 0.7 | 10.1 | 0.0 | |||||
| 10 | 3.6 | 2.1 | 5.7 | 0.0 | 9.0 | 0.8 | 9.8 | 0.0 | |||||
| n = 60 | 1 | 1.4 | 7.2 | 8.6 | 0.0 | 3.2 | 1.3 | 4.5 | 0.9 | 0.0 | 1.1 | 1.1 | 12.9 |
| 3 | 0.3 | 7.6 | 7.9 | 0.0 | 3.8 | 2.8 | 6.6 | 0.0 | 5.4 | 1.1 | 6.5 | 0.0 | |
| 5 | 0.5 | 7.0 | 7.5 | 0.0 | 3.1 | 2.7 | 5.8 | 0.0 | 4.9 | 0.9 | 5.8 | 0.0 | |
| 10 | 0.8 | 7.6 | 8.4 | 0.0 | 2.2 | 2.6 | 4.8 | 0.0 | 4.8 | 2.2 | 7.0 | 0.0 | |
L(U) is the proportion of lower (upper) endpoints of 95% confidence intervals that are greater (less) than the true Ne of 20, 50, or 100. “Tot” is the proportion of all confidence intervals that do not contain the true Ne. N̂ȯe is the proportion of simulated data sets for which the maximum-likelihood estimate of Ne was >400 (and for which the confidence interval was not considered). Values are from 1000 simulated data sets of n diploids sampled T generations apart. Following Tallmon et al. (2004) simulations of n = 20 when Ne = 100 were not done.