Sequence variants selected from a multi-breed GWAS can improve the reliability of genomic predictions in dairy cattle

Table 6 Differences in reliability between the WB-50 K model and other models for multi-breed prediction

Breed^a	Trait	MB-50 K (Δ^b)	MB-50 K + MBQTLt (Δ ^c_max )	MB-50 K + MBQTLt-n/w (Δ ^c_max )
HOLDK	Milk	0.07	0.08	0.09
	Fat	0.06	0.09	0.10
	Protein	0.04	0.05	0.06
HOLFR	Milk	0.05	0.08	0.08
	Fat	0.06	0.09	0.10
	Protein	0.04	0.06	0.06
JER	Milk	−0.03	0.02	0.06
	Fat	0.01	−0.01	0.04
	Protein	−0.02	0.02	0.03
RDC	Milk	0.01	0.05	0.06
	Fat	0.02	0.06	0.05
	Protein	0.00	0.02	0.04

The model included one genetic component with all 50 K SNPs (MB-50 K), or an additional component that included either all SNPs selected with a p value less than a threshold for the multi-breed GWAS (MB-50 K + MBQTLt) or a limited number of variants per QTL interval selected with a p value less than a threshold for the multi-breed GWAS (MB-50 K + MBQTLt-n/w)
^aHOLDK: Danish Holstein, HOLFR: French Holstein, JER: Jersey, RDC: Danish Red
^bDifferences in reliability between the reliability obtained with WB-50 K and the other models (Δ)
^cFor the models with a QTL component, the table provides only the Δ_max, i.e. the Δ obtained with the QTL set that resulted in the highest reliability

ISSN: 1297-9686