Genomic prediction using imputed whole-genome sequence data in Holstein Friesian cattle

Table 2 Estimated prediction reliability per pedigree group

Trait	Genotype data	Method	SMGS^a	SIRE^b (% of SMGS)	GS^c (% of SMGS)
SCS	Pedigree	BLUP	0.35	0.33 (94 %)	0.23 (67 %)
	BovineHD	GBLUP	0.53	0.50 (95 %)	0.45 (85 %)
	BovineHD	BSSVS	0.53	0.51 (95 %)	0.46 (86 %)
	ImputedHD	GBLUP	0.51	0.52 (101 %)	0.42 (83 %)
	ImputedHD	BSSVS	0.52	0.52 (102 %)	0.44 (85 %)
	Sequence	GBLUP	0.50	0.53 (104 %)	0.43 (85 %)
	Sequence	BSSVS	0.51	0.53 (103 %)	0.44 (87 %)
IFL	Pedigree	BLUP	0.29	0.16 (55 %)	0.15 (51 %)
	BovineHD	GBLUP	0.40	0.34 (85 %)	0.30 (75 %)
	BovineHD	BSSVS	0.42	0.34 (80 %)	0.31 (74 %)
	ImputedHD	GBLUP	0.39	0.32 (81 %)	0.25 (65 %)
	ImputedHD	BSSVS	0.41	0.31 (75 %)	0.27 (65 %)
	Sequence	GBLUP	0.39	0.32 (83 %)	0.25 (65 %)
	Sequence	BSSVS	0.41	0.32 (78 %)	0.27 (65 %)
PY	Pedigree	BLUP	0.30	0.30 (101 %)	0.24 (81 %)
	BovineHD	GBLUP	0.48	0.48 (100 %)	0.45 (95 %)
	BovineHD	BSSVS	0.49	0.49 (101 %)	0.45 (91 %)
	ImputedHD	GBLUP	0.45	0.43 (96 %)	0.41 (91 %)
	ImputedHD	BSSVS	0.47	0.47 (100 %)	0.41 (88 %)
	Sequence	GBLUP	0.45	0.45 (99 %)	0.42 (93 %)
	Sequence	BSSVS	0.46	0.45 (98 %)	0.42 (90 %)

Estimates of prediction reliability for somatic cell score (SCS), interval between first and last insemination (IFL), and protein yield (PY). Validation animals were divided based on the presence of relatives in the training set: sire and maternal grandsire (SMGS); only sire (SIRE); no sire, but one or two grandsires (GS). ^aStandard error of prediction reliability for the SMGS set was 0.02; ^bstandard error of prediction reliability for the SIRE set ranged from 0.06 to 0.08; ^cstandard error of prediction reliability for the SMGS set ranged from 0.03 to 0.05

ISSN: 1297-9686