TABLE 1.
Performance of single-locus and multilocus methods in mapping two QTL in 1000 simulated F18 advanced intercross populations
| Single locus or multilocus | Threshold
|
False discovery rate (%)e | Chromosome 2 associations (%)f | ||||
|---|---|---|---|---|---|---|---|
| Methoda | Typeb | Units | Valuec | Power (%)d | |||
| Naive single locus | Single | Permutation | Log P | 3.01 (2.67, 3.38) | 94.8 (±0.49) | 65.4 (±0.66) | 9.68 (±0.13) |
| Naive single locus | Single | Parametric bootstrap | Log P | 6.79 (3.01, 12.4) | 70.1 (±1.1) | 21.1 (±0.83) | 1.3 (±0.051) |
| Hard correction | Single | Permutation | Log P | 3.02 (2.66, 3.4) | 0.8 (±0.2) | 0.05 (±0.05) | 0 (±0) |
| Soft correction | Single | Permutation | Log P | 3.01 (2.65, 3.39) | 28.2 (±1) | 1.75 (±0.3) | 0 (±0) |
| Pedigree correction | Single | Permutation | Log P | 3.01 (2.63, 3.42) | 25 (±1) | 1.92 (±0.32) | 0 (±0) |
| Mixed model | Single | Parametric bootstrap | Log P | 3.23 (2.83, 3.63) | 83 (±0.86) | 26 (±0.82) | 0.318 (±0.025) |
| Single locus | Single | ∼80% power | Log P | 5.27 | 80 (±0.99) | 37.6 (±0.96) | 2.38 (±0.069) |
| Hard correction | Single | ∼80% power | Log P | 0.536 | 80 (±0.9) | 62.5 (±0.65) | 4.63 (±0.095) |
| Soft correction | Single | ∼80% power | Log P | 1.68 | 80 (±0.89) | 23.4 (±0.81) | 0.259 (±0.023) |
| Pedigree correction | Single | ∼80% power | Log P | 1.55 | 80 (±0.91) | 24.2 (±0.8) | 0.29 (±0.024) |
| Mixed model | Single | ∼80% power | Log P | 3.45 | 80 (±0.92) | 22.3 (±0.81) | 0.216 (±0.021) |
| Bagging | Multilocus | ∼80% power | Range probability | 0.65 | 80.6 (±0.93) | 1.83 (±0.36) | 0.00204 (±0.002) |
| Subagging (40%) | Multilocus | ∼80% power | Range probability | 0.6 | 80.1 (±0.93) | 1.6 (±0.35) | 0.00204 (±0.002) |
| Subagging (50%) | Multilocus | ∼80% power | Range probability | 0.65 | 80.5 (±0.94) | 1.52 (±0.33) | 0.00204 (±0.002) |
| Subagging (60%) | Multilocus | ∼80% power | Range probability | 0.7 | 80.6 (±0.93) | 1.82 (±0.37) | 0.00612 (±0.0035) |
| Subagging (70%) | Multilocus | ∼80% power | Range probability | 0.76 | 80.2 (±0.95) | 1.78 (±0.37) | 0.0122 (±0.005) |
| Subagging (80%) | Multilocus | ∼80% power | Range probability | 0.82 | 80.4 (±0.94) | 2.35 (±0.42) | 0.0143 (±0.0054) |
| Subagging (90%) | Multilocus | ∼80% power | Range probability | 0.91 | 80.2 (±0.96) | 2.72 (±0.45) | 0.0143 (±0.0054) |
| Subagging (80%, medium stopping rule) | Multilocus | ∼80% power | Range probability | 0.06 | 81.6 (±1) | 5.78 (±0.54) | 0.939 (±0.044) |
| Subagging (80%, strict stopping rule) | Multilocus | ∼80% power | Range probability | 0 | 100 (±0) | 90 (±0) | 100 (±0) |
| Forward selection | Multilocus | NA | NA | NA | 90.4 (±0.71) | 8.51 (±0.66) | 0.147 (±0.017) |
| Bagging | Multilocus | ∼90.35% power | Range probability | 0.46 | 90.5 (±0.69) | 5.73 (±0.55) | 0.0224 (±0.0068) |
| Subagging (80%) | Multilocus | ∼90.35% power | Range probability | 0.5 | 90.4 (±0.71) | 6.72 (±0.6) | 0.0571 (±0.011) |
Subagging (p%) refers to resample model averaging using p% subsamples. “Strict stopping” refers to use of a stopping rule based on thresholds from parametric bootstrap of a multilevel sibship model; “medium stopping” refers to using the mean of that threshold and one derived by permutation; all other stopping rules are based on permutation.
How the threshold for calculating performance statistics was determined and applied. Permutation: genomewide 5% significance thresholds were calculated separately for each simulation. Parametric bootstrap: thresholds were calculated separately for each simulation. “∼n% power”: scores were pooled for all simulations and the threshold was calibrated to give ∼n% power.
For the permutation and parametric bootstrap results, threshold values are given as “median (lowest, highest).” Otherwise the numbers are the range probability thresholds required to achieve ∼n% power.
The proportion of segments containing QTL that had scores above the threshold, averaged over simulations (SE in parentheses).
The proportion of 10-cM segments with scores above the threshold that did not contain a QTL, averaged over simulations (SE in parentheses).
The proportion of marker intervals on chromosome 2 predicted to contain QTL, averaged over simulations (SE in parentheses).