Skip to main content
. 2019 Jan 10;122(6):848–863. doi: 10.1038/s41437-018-0172-0

Table 3.

The corresponding predictive abilities for GS analyses in Table 2, calculated as the correlation between the raw phenotype (juvenile height: HTJ) and their genomic estimated breeding values (GEBVs) or deregressed genomic estimated breeding values (GDEBVs)

Analysis Predictive ability (SE)
ABLUP RR-BLUP GRR Bayes-B
r (HTJ, HTJ GEBVs) 0.71 (0.003) 0.43 (0.0004) 0.42 (0.0006) 0.43 (0.0007)
EBV < 201 0.10 (0.002) 0.07 (0.002) 0.07 (0.003)
EBV < 202 0.04 (0.002) −0.01 (0.003) −0.01 (0.003)
EBV < 203 −0.10 (0.003) −0.12 (0.004) −0.09 (0.003)
EBV > 201 0.05 (0.002) 0.03 (0.002) 0.07 (0.004)
EBV > 202 0.04 (0.002) 0.01 (0.002) 0.07 (0.003)
EBV > 203 0.15 (0.004) 0.12 (0.005) 0.59 (0.0004)
r (HTJ, HTJ GDEBVs) 0.06 (0.008) 0.009 (0.007) 0.40 (0.003)
EBV < 201 −0.20 (0.006) −0.21 (0.005) −0.04 (0.006)
EBV < 202 −0.22 (0.005) −0.25 (0.005) −0.15 (0.007)
EBV < 203 −0.21 (0.006) −0.20 (0.006) −0.21 (0.006)
EBV > 201 0.23 (0.004) 0.31 (0.005) 0.20 (0.005)
EBV > 202 0.19 (0.005) 0.22 (0.005) 0.16 (0.006)
EBV > 203 0.15 (0.007) 0.11 (0.007) 0.07 (0.010)
r (HTJ, HT35 GEBVs) 0.24 (0.005) 0.21 (0.001) 0.21 (0.001) 0.22 (0.002)
EBV < 201 −0.19 (0.001) −0.13 (0.002) −0.09 (0.005)
EBV < 202 −0.23 (0.001) −0.17 (0.002) −0.14 (0.004)
EBV < 203 −0.10 (0.002) −0.07 (0.002) −0.06 (0.004)
EBV > 201 −0.03 (0.001) −0.04 (0.001) −0.09 (0.002)
EBV > 202 −0.03 (0.0009) −0.05 (0.001) −0.07 (0.002)
EBV > 203 −0.04 (0.001) −0.04 (0.002) −0.06 (0.001)
r (HTJ, HT35 GDEBVs) −0.17 (0.005) −0.13 (0.004) 0.04 (0.006)
EBV < 201 −0.18 (0.003) −0.13 (0.003) −0.13 (0.003)
EBV < 202 −0.18 (0.003) −0.18 (0.003) −0.18 (0.003)
EBV < 203 −0.05 (0.003) −0.03 (0.004) −0.05 (0.004)
EBV > 201 0.08 (0.005) 0.13 (0.004) 0.003 (0.004)
EBV > 202 0.09 (0.005) 0.09 (0.006) −0.01 (0.003)
EBV > 203 0.15 (0.005) 0.15 (0.005) 0.06 (0.006)

Results for the validation set as a whole are in bold (N = 136), below each are the results for each of the two clusters EBV < 20 (N = 83) and EBV > 20 (N = 53), with indices representing different training set composition: 1 all genotyped F1 individuals; 2 all genotyped F1 individuals minus the parents of the opposing cluster; 3 only the F1 parents of the cluster in question