Using imputation-based whole-genome sequencing data to improve the accuracy of genomic prediction for combined populations in pigs

Table 4 Accuracy of genomic prediction for production traits using different reference populations to predict 270 younger individuals of the LM population

Reference set^a (size of reference population)	Method	Trait
Reference set^a (size of reference population)	Method	AGE	BFT
LM (1499)	GBLUP_80K	0.625	0.432
	GBLUP_WGS	0.633	0.436
	GBLUP_LD	0.642	0.445
	GFBLUP_QTL	0.642	0.444
	GFBLUP_GWAS	0.640	0.444
	ssGBLUP_80K	0.739	0.677
	ssGBLUP_WGS	0.738	0.663
	ssGBLUP_LD	0.745	0.671
LM+ZX (1819)	GBLUP_80K	0.615	0.431
	GBLUP_WGS	0.636	0.443
	GBLUP_LD	0.648	0.454
	GFBLUP_QTL	0.638	0.443
	GFBLUP_GWAS	0.635	0.452
	ssGBLUP_80K	0.737	0.677
	ssGBLUP_WGS	0.739	0.664
	ssGBLUP_LD	0.746	0.673
ZX (320)	GBLUP_80K	0.028	0.067
	GBLUP_WGS	0.098	0.111
	GBLUP_LD	0.164	0.142
	GFBLUP_QTL	0.164	0.144
	GFBLUP_GWAS	0.124	0.119

GBLUP_80K genomic BLUP based on the 80K SNP panel; GBLUP_WGS GBLUP based on imputed whole-genome sequencing (WGS) data; GBLUP_LD GBLUP based on WGS after LD pruning; GFBLUP_QTL genomic feature BLUP with a genomic feature matrix constructed based on QTL prior information; GFBLUP_GWAS genomic feature BLUP with a genomic feature matrix constructed based on GWAS prior information; ssGBLUP_80K, ssGBLUP_WGS, ssGBLUP_LD For ssGBLUP, the H matrix was constructed based on the different genomic relationship matrices of the 80K chip panel, WGS data and WGS data after LD pruning and termed ssGBLUP_80K, ssGBLUP_WGS and ssGBLUP_LD, respectively; AGE days to 100 kg; BFT backfat thickness
^aYorkshire population of LM, ZX or LM plus the ZX reference population

ISSN: 1297-9686