Effect of genomic selection and genotyping strategy on estimation of variance components in animal models using different relationship matrices

Table 8 Average bias of predictions^a from H-AM with different variance components (VC) in the simulated data

Scenario	Genotyping strategy (%)		VC-H		VC-A		VC-bp
Scenario	Genotyping strategy (%)		Genotyped	Ungenotyped	Genotyped	Ungenotyped	Genotyped	Ungenotyped
1	Selective	10	0.96	0.76	0.96	0.82	0.96	0.88
		20	0.93	0.33	0.99	0.85	0.99	0.91
		30	0.97	0.26	1.04	0.88	1.04	0.93
	Random	10	0.91	0.87	0.88	0.79	0.91	0.86
		20	0.94	0.89	0.91	0.80	0.93	0.87
		30	0.95	0.90	0.93	0.81	0.95	0.88
2	Selective	10	1.01	0.80	1.02	0.92	1.02	0.93
		20	0.98	0.31	1.06	0.95	1.06	0.96
		30	1.01	0.27	1.10	0.98	1.10	0.98
	Random	10	0.96	0.95	0.94	0.91	0.95	0.92
		20	0.99	0.96	0.97	0.91	0.97	0.92
		30	0.98	0.96	0.97	0.92	0.97	0.93
3	Selective	10	1.01	0.91	1.03	1.00	1.02	0.97
		20	0.99	0.41	1.07	1.02	1.06	0.99
		30	1.06	0.31	1.11	1.03	1.11	1.01
	Random	10	0.99	0.99	0.99	1.00	0.97	0.95
		20	0.99	0.99	0.99	1.00	0.98	0.96
		30	0.99	1.00	0.99	1.00	0.98	0.96

H-AM animal model with combined pedigree-based and genomic relationship matrix, VC-H VC estimates from H-AM, VC-A VC estimates from A-AM (animal model with pedigree-based relationship matrix), VC-bp VC of the base population
^aBias of predicted breeding values is presented as the regression coefficient of the true breeding values on the predicted breeding values; deviation from 1 implies bias, regression coefficients lower than 1 imply inflation in the predictions

ISSN: 1297-9686