Fig. 1.
Fig. 1.

Mean of the variance components (and their standard deviation) of the marker main effect and the residual for each of the four environments (1–4) in the wheat data set for the genomic best linear unbiased prediction (GBLUP), reproducing kernel Hilbert space (RKHS) empirical Bayes (EB), and RKHS kernel averaging (KA) models.

 


Fig. 2.Fig. 2.Fig. 2.
Fig. 2.

(continued on next page) For six pairs of environments in the wheat data sets (1–2, 1–3, 1–4, 2–3, 2–4, and 3–4), the plot shows the mean of the variance components for three models of the (A) residual variance component, (B) marker main effect, and (C) marker environment specific effect. Marker-specific effect 1 denotes the variance component of the marker effect associated with the first environment in the pair (e.g., 3 in pair 3–4), whereas marker-specific effect 2 is related to the variance component of the second environment in that pair (e.g., 4 in pair 3–4).

 


Fig. 3.
Fig. 3.

Mean of the variance components and their standard deviation of the marker main effect and the residual for each of the three environments (1–3) in the maize data set for the genomic best linear unbiased prediction (GBLUP), reproducing kernel Hilbert space (RKHS) empirical Bayes (EB), and RKHS kernel averaging (KA) models.

 


Fig. 4.Fig. 4.Fig. 4.
Fig. 4.

(continued on next page) For three pairs of environments in the maize data sets (1–2, 1–3, 2–3, and 2–3), the plot shows the mean of the variance components for three models (genomic best linear unbiased prediction [GBLUP]–genotype × environment interaction [G × E] model, reproducing kernel Hilbert space [RKHS] empirical Bayes [EB]–G × E, and RKHS kernel averaging [KA]–G × E [and their standard deviation]) of the (A) residual variance component, (B) marker main effect, and (C) marker environment specific effect. Marker-specific effect 1 denotes the variance component of the marker effect associated with the first environment in the pair (e.g., 2 in pair 2–3), whereas marker-specific effect 2 is related to the variance component of the second environment in that pair (e.g., 3 in pair 2–3).

 


Box A1.
Box A1.

Estimation of the bandwidth (h) parameter (adapted from Pérez-Elizalde et al., 2015).

 


Box A2.
Box A2.

Function to generate the missing values for cross-validation 2 design (CV2) scheme (López-Cruz et al., 2015).

 


Box A3.
Box A3.

Function to fit genotype × environment interaction model (adapted from López-Cruz et al., 2015).

 


Box A4.
Box A4.

An example of fitting the reproducing kernel Hilbert space (RKHS) empirical Bayes (EB)–genotype × environment interaction model.

 


Box A5.
Box A5.

Fitting the reproducing kernel Hilbert space (RKHS) kernel averaging (KA)–genotype × environment interaction model.

 


Box A6.
Box A6.

An example of reproducing kernel Hilbert space (RKHS) kernel averaging (KA)–genotype × environment interaction model.