Table 5.
Inflation* of genomic predictions from different methods, obtained for five traits of a mice population
| Splitting | Method |
Trait |
||||
|---|---|---|---|---|---|---|
| W6W | WGS | BL | %CD8+ | CD4+/CD8+ | ||
| Within |
POL |
1.21 |
0.91 |
0.89 |
1.03 |
0.97 |
| emBayesB |
1.43 |
1.57 |
3.70 |
1.65 |
1.55 |
|
| RR_GBLUP |
0.95 |
0.70 |
0.88 |
0.66 |
0.75 |
|
| SS_BY |
0.79 |
0.66 |
- |
0.60 |
0.72 |
|
| SS_ABS |
0.72 |
0.49 |
0.49 |
0.64 |
0.72 |
|
| RKHS |
1.08 |
0.95 |
0.89 |
1.06 |
1.26 |
|
| SVR |
1.24 |
0.91 |
0.68 |
1.16 |
1.18 |
|
| BayesCpi |
0.55 |
0.29 |
0.40 |
0.17 |
0.29 |
|
| BayesC |
0.18 |
0.12 |
0.21 |
0.19 |
0.20 |
|
| LASSO |
0.93 |
1.01 |
1.08 |
1.00 |
1.07 |
|
| RF |
1.48 |
1.10 |
0.78 |
1.12 |
1.11 |
|
| Across | POL |
- |
- |
- |
- |
- |
| emBayesB |
0.46 |
0.90 |
1.47 |
1.43 |
1.33 |
|
| RR_GBLUP |
0.37 |
0.49 |
0.52 |
0.46 |
0.52 |
|
| SS_BY |
0.30 |
0.46 |
- |
0.42 |
0.52 |
|
| SS_ABS |
0.29 |
0.30 |
0.32 |
0.50 |
0.52 |
|
| RKHS |
0.55 |
0.95 |
0.75 |
1.21 |
1.36 |
|
| SVR |
0.61 |
0.89 |
0.62 |
1.42 |
1.27 |
|
| BayesCpi |
0.22 |
0.21 |
0.14 |
0.17 |
0.24 |
|
| BayesC |
0.07 |
0.08 |
0.11 |
0.18 |
0.15 |
|
| LASSO |
0.38 |
0.71 |
0.56 |
0.98 |
1.01 |
|
| RF | 0.72 | 1.22 | 0.65 | 1.27 | 1.22 | |
*Average of ten replicates. Inflation was measured as the slope of the regression of observed phenotypes on predicted phenotypes of testing set
Trait = weight at 6 weeks (W6W), weight growth slope (WGS), body length (BL), percentage of CD8+ cells (%CD8+), ratio between CD4+ and CD8+ cells (CD4+/CD8+). Splitting= splitting strategy in cross-validation (within or across-family).