TABLE 6.
Bias of GW-EBVs obtained by G-BLUP, MIXTURE, and BayesB
| G-BLUP | MIXTURE | BayesB | |
|---|---|---|---|
| Cohort maskinga | |||
| Milk yield | 1.037 | 1.047 | 1.008 |
| Fat yield | 1.075 | 1.119 | 1.020 |
| Protein yield | 1.080 | 1.106 | 1.044 |
| cm1 | 0.693 | 0.849 | 0.502 |
| cm2 | 0.550 | 0.546 | 0.323 |
| cm3 | 0.718 | 0.831 | 0.475 |
| Calving ease | 1.192 | 0.932 | 0.998 |
| Random maskinga | |||
| Milk yield | 0.984 | 0.995 | 0.951 |
| Fat yield | 1.034 | 1.056 | 0.967 |
| Protein yield | 1.046 | 1.075 | 0.992 |
| cm1 | 0.628 | 0.842 | 0.467 |
| cm2 | 0.693 | 0.814 | 0.466 |
| cm3 | 0.771 | 0.921 | 0.525 |
| Calving ease | 1.104 | 0.953 | 0.862 |
| Approximate SE |
0.089 |
0.119 |
0.070 |
a
The bias for cohort masking is shown as a combined bias estimated for 500 selected individuals, while the bias for random masking is shown as the mean of the biases for five training data sets and the pooled approximate standard error.