Skip to main content

Table 6 Normalized root-mean squared error(NRMSE) * of genomic predictions from different methods, obtained for five traits of a mice population

From: A comparison of statistical methods for genomic selection in a mice population

Splitting Method Trait
W6W WGS BL %CD8+ CD4+/CD8+
Within POL 0.099 0.114 0.123 0.138 0.145
emBayesB 0.140 0.122 0.124 0.165 0.165
RR_GBLUP 0.103 0.114 0.123 0.147 0.136
SS_BY 0.112 0.122 - 0.158 0.144
SS_ABS 0.117 0.120 0.132 0.161 0.143
RKHS 0.098 0.112 0.122 0.127 0.131
SVR 0.099 0.112 0.123 0.128 0.130
BayesCpi 0.166 0.214 0.268 1.252 0.502
BayesC 0.534 0.877 0.251 1.035 1.080
LASSO 0.172 0.137 0.127 0.204 0.196
RF 0.102 0.112 0.123 0.124 0.119
Across POL - - - - -
emBayesB 0.154 0.132 0.126 0.199 0.230
RR_GBLUP 0.128 0.121 0.125 0.184 0.170
SS_BY 0.141 0.126 - 0.193 0.175
SS_ABS 0.145 0.130 0.140 0.188 0.180
RKHS 0.122 0.119 0.124 0.152 0.152
SVR 0.122 0.119 0.125 0.155 0.153
BayesCpi 0.163 0.243 0.240 1.519 0.593
BayesC 0.637 0.747 0.275 0.896 1.091
LASSO 0.201 0.142 0.129 0.213 0.225
RF 0.119 0.119 0.124 0.144 0.130
  1. *Average of ten replicates. Lower values are associated with better overall fit
  2. 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).