Table 4.

Predictabilities of the five traits based on MultiBLUP framework using various combinations of the genetic relatedness matrix

Trait	mBLUP (2)a	mBLUP (4)	mBLUP (adapt)	GBLUP (whole)
YD	0.451 ± 0.010b	0.451 ± 0.008	0.445 ± 0.015	0.451 ± 0.008
HD	0.560 ± 0.008	0.556 ± 0.011	0.550 ± 0.004	0.555 ± 0.011
PH	0.701 ± 0.005	0.696 ± 0.007	0.695 ± 0.009	0.693 ± 0.007
GL	0.553 ± 0.004	0.550 ± 0.008	0.576 ± 0.005	0.549 ± 0.008
GW	0.696 ± 0.004	0.697 ± 0.006	0.695 ± 0.008	0.694 ± 0.006

^aDenotes the number of individual genetic relatedness, such as 2 representing the two relatedness matrices was used, “adapt” representing the number of matrices automatically calculated by LDAK software and “whole” denotes that whole genetic variants were incorporated into one matrix

^bDenotes the standard error of predictability of traits, calculated from 100 replicates of tenfold cross-validation