Table 3 Optimal sampling observed for individual datasets to enhance predictive ability (PA) and mean squared prediction error (MSPE). Subsampling performed with (wR) and without replacement (woR)
Optimal PA | Increase in PA | Optimal MSPE | Decrease in MSPE | |
|---|---|---|---|---|
Mouse | wR 66% | 2.5% | woR 32% | <0.1% |
Soybean | woR 25% | 0.1% | woR 25% | 0.1% |
Wheat | woR 34% | 0.7% | woR 33% | 0.5% |
Simulated F2 | wR 87% | 0.1% | wR 66% | 0.2% |