Single-step genomic evaluation with metafounders for feed conversion ratio and average daily gain in Danish Landrace and Yorkshire pigs

Table 4 Accuracies of genomic predictions for average daily gain (ADG) and feed conversion ratio (FCR) in different validation groups for univariate (UNI) or bivariate (BI) single-step genomic prediction models with or without metafounders (MF), each with their optimal weighting factor (ω) in the two populations

Population	Trait	Validation group	UNI_SSGBLUP	UNI_MF-SSGBLUP	BI_SSGBLUP	BI_MF-SSGBLUP
Landrace	ADG (ω = 0.20)	All	0.355^a	0.356^a	0.355^a	0.356^a
		Genotyped	0.410^a	0.410^a	0.410^a	0.410^a
		Non-genotyped	0.327^a	0.327^a	0.326^a	0.327^a
	FCR (ω = 0.70)	All	0.192^a	0.193^a	0.193^a	0.196^b
		Genotyped	0.200^a	0.206^b	0.201^a	0.206^b
		Non-genotyped	0.167^a	0.170^a	0.179^b	0.180^b
Yorkshire	ADG (ω = 0.25)	All	0.347^a	0.347^a	0.347^a	0.347^a
		Genotyped	0.395^a	0.395^a	0.394^a	0.394^a
		Non-genotyped	0.309^a	0.309^a	0.309^a	0.308^a
	FCR (ω = 0.10)	All	0.355^a	0.355^a	0.344^b	0.344^b
		Genotyped	0.370^a	0.367^a	0.357^b	0.356^b
		Non-genotyped	0.118^a	0.121^a	0.117^a	0.114^a

The four methods are UNI_SSGBLUP, UNI_MF-SSGBLUP, BI_SSGBLUP and BI_MF-SSGBLUP
Validation groups are the validation population including all the young pigs born between 1 May 2016 and 1 February 2017, and two subgroups of genotyped and non-genotyped animals
The results for ADG and FCR bivariate single-step genomic prediction models are obtained with their optimal ω respectively
Superscript letters indicate significant differences (p < 0.05) by the Hotelling–Williams t-test

ISSN: 1297-9686