Table 3. Mean predictive ability of genomic selection models using tetraploid dosage (GS-TD) and diploid dosage (GS-DD) for organic matter (OM), crude protein (CP), in vitro digestibility of organic matter (IVD), leaf dry matter (LDM), regrowth capacity (RC), and percentage of leaf blade (PLB). GS-TD and GS-DD models with the highest mean predictive ability for each trait are indicated in bold. Molecular matrices containing 1,223 markers for each data set. The broad-sense heritability is ${H^}_C^2$ and the narrow-sense heritability is ${h^}^2$.

Model	OM	CP	IVD	LDM	RC	PLB
GBLUP-TD	0.3782	0.2237	0.2211	0.1282	0.2475	0.1957
GBLUP-DD	0.2905	0.1578	0.1869	0.0685	0.1829	0.1968
BRR-TD	0.3866	0.2299	0.2157	0.1390	0.2621	0.2058
BRR-DD	0.2966	0.1633	0.1837	0.0997	0.1951	0.2155
BA-TD	0.3931	0.2251	0.2174	0.1308	0.2514	0.2001
BA-DD	0.2958	0.1585	0.1835	0.0854	0.1876	0.2067
BB-TD	0.3955	0.2242	0.2190	0.1311	0.2501	0.2004
BB-DD	0.2960	0.1560	0.1843	0.0886	0.1885	0.2110
BC-TD	0.3879	0.2282	0.2186	0.1374	0.2577	0.2049
BC-DD	0.2958	0.1612	0.1855	0.1006	0.1944	0.2161
BL-TD	0.3821	0.2233	0.2187	0.1333	0.2468	0.1991
BL-DD	0.2923	0.1594	0.1849	0.0870	0.1853	0.1998
Average GS-TD	0.3872	0.2257	0.2184	0.1333	0.2526	0.2010
Average GS-DD	0.2945	0.1594	0.1848	0.0883	0.1889	0.2076
Average	31.48	41.59	18.18	50.96	33.72	−3.18
superiority
${\widehat{H}}_C^2$	0.66	0.42	0.48	0.89	0.85	0.43
${\widehat{h}}^2$	0.53	0.23	0.25	0.11	0.31	0.12