Bias in estimates of variance components in populations undergoing genomic selection: a simulation study

BMC Genomics

Table 2 Mean (SD) of the “true” variance components and heritability in the base (founder) population (P-base) and the base population for the genomic phase^a and estimates of variance components and heritabilities from P-AM, ssGBLUP, and ssBR based on three scenarios of phenotyping^b

Method	Scenario^b	\( {\sigma}_g^2 \)(\( {\sigma}_{\varepsilon}^2 \))	\( {\sigma}_e^2 \)	\( {\sigma}_{\alpha}^2\sum \limits_{j=1}^m2{p}_j\left(1-{p}_j\right) \)	h^2c
True VCs in P-base		3.59 (0.03)	5.05 (0.09)	–	0.42 (0.004)
True VCs in G-base		2.82 (0.09)	5.10 (0.07)	–	0.36 (0.009)
P-AM	Pheno₁	3.62 (0.12)	5.02 (0.02)	–	0.42 (0.009)
	Pheno_1 + 2	3.52 (0.06)	5.07 (0.03)	–	0.41 (0.004)
	Pheno₂	3.19 (0.20)^*	5.21 (0.10)	–	0.38 (0.018)^*
ssGBLUP	Pheno₁	2.99 (0.12)^***	5.54 (0.01)^***	–	0.35 (0.009)^***
	Pheno_1 + 2	3.39 (0.06)^***	5.17 (0.03)	–	0.40 (0.009)^***
	Pheno₂	3.36 (0.15)^***	5.19 (0.05)	–	0.39 (0.013)^*
ssBR	Pheno₁	3.44 (0.15)	5.10 (0.04)	4.20 (0.10)^***	0.40 (0.012)
	Pheno_1 + 2	3.14 (0.10)^***	5.24 (0.02)^*	3.71 (0.11)	0.37 (0.008)^***
	Pheno₂	2.99 (0.20)	5.27 (0.06)^**	2.97 (0.09)	0.36 (0.017)

^aP-AM: traditional pedigree-based animal model; ssGBLUP: single-step genomic BLUP; ssBR: single-step Bayesian regression. \( {\sigma}_g^2 \) is the genetic variance used in P-AM, \( {\sigma}_{\varepsilon}^2 \) is the total genetic variance used in ssBR; \( {\sigma}_e^2 \) is the residual variance; \( {\sigma}_{\alpha}^2 \) is the marker variance; \( {\sigma}_{\alpha}^2\sum \limits_{j=1}^m2{p}_j\left(1-{p}_j\right) \) is used to calculate genetic variance via the estimated marker variance in ssBR, where p_j is the observed allele frequency at locus j, and m is the total number of markers
^bPheno₁: phenotypes from only the conventional phase (1–20 years) were used; Pheno_1 + 2: phenotypes from both the conventional phase and genomic selection phase (1–35 years) were used; Pheno₂: phenotypes from only the genomic selection phase were used (21–35 years)
^cHeritabilities (h²) from P-AM, ssGBLUP, and ssBR calculated as \( \frac{\sigma_{g\left(\varepsilon \right)}^2}{\left({\sigma}_{g\left(\varepsilon \right)}^2+{\sigma}_e^2\right)} \)
The significance test was performed to determine whether the estimated parameter differs from the simulated parameter. ^* significant at P < 0.01; ^** significant at P < 0.005; ^*** significant at P < 0.001

ISSN: 1471-2164