TABLE 4.

Genome-wide prediction abilities of flowering time (FT), plant height (PH), and thousand grain weight (TGW) obtained by applying five-fold cross validation to the entire set of phenotyped spring barley accessions modeling the row type as covariate (RT).

Trait	Method	All	2-Rowed	6-Rowed	Deficiens	Intermedium	Labile
FT	RT-covariatea	0.734 (0.003)	0.685 (0.004)	0.741 (0.003)	0.638 (0.009)	0.779 (0.007)	0.859 (0.004)
	RT-ignoredb	0.734 (0.003)	0.685 (0.004)	0.741 (0.004)	0.638 (0.008)	0.778 (0.006)	0.859 (0.004)
	RT-withinc	n.a.	0.690 (0.003)	0.741 (0.003)	0.572 (0.032)	0.690 (0.036)	0.861 (0.008)
PH	RT-covariate	0.794 (0.001)	0.722 (0.002)	0.817 (0.001)	0.623 (0.008)	0.749 (0.007)	0.514 (0.010)
	RT-ignored	0.791 (0.001)	0.719 (0.002)	0.815 (0.001)	0.614 (0.008)	0.751 (0.007)	0.530 (0.011)
	RT-within	n.a.	0.718 (0.003)	0.815 (0.001)	0.601 (0.013)	0.670 (0.011)	0.544 (0.023)
TGW	RT-covariate	0.855 (0.001)	0.719 (0.004)	0.862 (0.001)	0.650 (0.012)	0.849 (0.006)	0.425 (0.019)
	RT-ignored	0.841 (0.001)	0.693 (0.005)	0.850 (0.002)	0.600 (0.011)	0.893 (0.005)	0.349 (0.014)
	RT-within	n.a.	0.712 (0.003)	0.863 (0.002)	0.633 (0.014)	0.852 (0.008)	0.477 (0.028)

^aRow type was treated as a covariate (fixed effect). ^bRow type was not considered in the model. ^cWith-in row type prediction, results were extracted from Table 3 for comparison. For each row type subpopulation, the prediction ability was compared with the result obtained in the within-subpopulation prediction. The standard deviations of the prediction abilities were presented in brackets.