Table 2 Genomic selection analyses of four models using three GS statistical methods (RR-BLUP, GRR, and Bayes-B)

From: Genomic selection of juvenile height across a single-generational gap in Douglas-fir

Analysis

Accuracy (SE)

ABLUP

RR-BLUP

GRR

Bayes-B

HTJ EBVs → HTJ GEBVs

0.92 (0.001)

0.92 (0.0002)

0.91 (0.0003)

0.91 (0.0007)

EBV < 201

 

0.43 (0.003)

0.42 (0.003)

0.42 (0.004)

EBV < 202

 

0.43 (0.004)

0.38 (0.005)

0.38 (0.005)

EBV < 203

 

0.45 (0.006)

0.43 (0.006)

0.44 (0.005)

EBV > 201

 

−0.005 (0.004)

−0.05 (0.004)

−0.02 (0.006)

EBV > 202

 

−0.04 (0.004)

−0.11 (0.004)

−0.10 (0.005)

EBV > 203

 

0.54 (0.009)

0.52 (0.010)

0.99 (0.0001)

HTJ DEBVS → HTJ GDEBVs

 

0.10 (0.008)

0.05 (0.007)

0.48 (0.002)

EBV < 201

 

−0.12 (0.004)

−0.11 (0.004)

−0.02 (0.005)

EBV < 202

 

−0.12 (0.004)

−0.14 (0.004)

−0.09 (0.005)

EBV < 203

 

−0.08 (0.004)

−0.07 (0.004)

−0.07 (0.004)

EBV > 201

 

0.17 (0.004)

0.26 (0.004)

0.16 (0.006)

EBV > 202

 

0.14 (0.004)

0.17 (0.004)

0.13 (0.006)

EBV > 203

 

−0.23 (0.003)

−0.22 (0.003)

0.09 (0.009)

HT35 EBVs → HTJ GEBVs

0.60 (0.010)

0.57 (0.002)

0.56 (0.002)

0.58 (0.003)

EBV < 201

 

0.15 (0.004)

0.20 (0.004)

0.04 (0.007)

EBV < 202

 

0.13 (0.004)

0.18 (0.004)

0.04 (0.007)

EBV < 203

 

0.35 (0.005)

0.41 (0.005)

0.36 (0.007)

EBV > 201

 

−0.21 (0.002)

−0.24 (0.002)

−0.28 (0.004)

EBV > 202

 

−0.22 (0.002)

−0.25 (0.002)

−0.30 (0.003)

EBV > 203

 

−0.23 (0.003)

−0.22 (0.003)

−0.27 (0.003)

HT35 DEBVs → HTJ GDEBVs

 

−0.15 (0.005)

−0.11 (0.005)

0.11 (0.007)

EBV < 201

 

−0.08 (0.003)

−0.02 (0.003)

−0.02 (0.003)

EBV < 202

 

−0.08 (0.003)

−0.07 (0.003)

−0.06 (0.003)

EBV < 203

 

0.06 (0.002)

0.07 (0.003)

0.06 (0.003)

EBV > 201

 

0.02 (0.004)

0.07 (0.004)

−0.06 (0.004)

EBV > 202

 

0.02 (0.005)

0.03 (0.006)

−0.07 (0.003)

EBV > 203

 

0.07 (0.004)

0.08 (0.004)

−0.01 (0.006)

  1. ABLUP is a pedigree only model with no marker information used. Results are from the validation procedure replicated 10 times, in which a random 90% of the genotyped F1 generation (1321 trees from Adams, Fleet River, and Lost Creek) was used as the training set and the validation set was comprised of the 136 genotyped F2 trees from Jordan River. Accuracy was calculated as the mean of the replications of the Pearson product-moment correlation between the original EBVs for HTJ of the 136 F2 validation trees from Jordan River and their predicted GEBVs or GDEBVs. The four analyses are: F1 juvenile height EBVs predicting F2 juvenile height EBVs (HTJ EBVs → HTJ GEBVs); F1 juvenile height DEBVs predicting F2 juvenile height DEBVs (HTJ DEBVS → HTJ GDEBVs); F1 mature (age 35) height EBVs predicting F2 juvenile height EBV (HT35 EBVs → HTJ GEBVs); and F1 mature height DEBVs predicting F2 juvenile height GDEBVs (HT35 DEBVs → HTJ GDEBVs). Results for the validation set as a whole are in bold (N = 136), following these 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
Back to article page