Weighted single-step genomic best linear unbiased prediction integrating variants selected from sequencing data by association and bioinformatics analyses

Liu, Aoxing; Lund, Mogens Sandø; Boichard, Didier; Karaman, Emre; Guldbrandtsen, Bernt; Fritz, Sebastien; Aamand, Gert Pedersen; Nielsen, Ulrik Sander; Sahana, Goutam; Wang, Yachun; Su, Guosheng

doi:10.1186/s12711-020-00568-0

Genetics Selection Evolution

Table 5 Regression coefficients of yield deviation (YD) on prediction from a pedigree-based BLUP (PBLUP) model and different ssGBLUP models using different SNP sets for milk production traits

From: Weighted single-step genomic best linear unbiased prediction integrating variants selected from sequencing data by association and bioinformatics analyses

Trait	Population	Model	Pedigree	SNP
Trait	Population	Model	Pedigree	54K	54K + DFS	54K + FRA	54K + DFS + FRA
Milk	Genotyped	PBLUP	0.89	–	–	–	–
		ssGBLUP	–	_a0.84^a	_a0.83^a	_a0.83^a	_a0.83^a
		ssFGBLUP	–	–	_a0.83^a	_b0.81^a	_a0.82^a
		ssWGBLUP_bin1	–	_c0.80^a	_b0.80^a	_c0.79^a	_b0.80^a
		ssWGBLUP_bin30	–	_b0.81^a	_b0.80^b	_bc0.80^b	_b0.80^b
		ssWGBLUP_bin100	–	_b0.81^a	_b0.81^ab	_c0.80^b	_b0.80^b
	Non-genotyped	PBLUP	0.87	–	–	–	–
		ssGBLUP	–	_a0.87^a	_a0.87^a	_a0.86^a	_a0.86^a
		ssFGBLUP	–	–	_ab0.87^a	_a0.86^b	_a0.86^ab
		ssWGBLUP_bin1	–	_b0.85^a	_c0.85^a	_b0.84^a	_b0.85^a
		ssWGBLUP_bin30	–	_ab0.85^a	_c0.85^a	_b0.85^a	_b0.85^a
		ssWGBLUP_bin100	–	_a0.86^a	_bc0.86^a	^ab0.85^ab	_b0.85^b
Fat	Genotyped	PBLUP	0.77	–	–	–	–
		ssGBLUP	–	_a0.74^a	_a0.74^a	_a0.73^a	_a0.73^a
		ssFGBLUP	–	–	_a0.75^a	_a0.72^b	_a0.73^b
		ssWGBLUP_bin1	–	_b0.68^a	_b0.68^a	_b0.68^a	_b0.68^a
		ssWGBLUP_bin30	–	_b0.69^a	_b0.70^a	_b0.69^a	_b0.70^a
		ssWGBLUP_bin100	–	_b0.70^a	_b0.69^a	_b0.69^a	_b0.68^a
	Non-genotyped	PBLUP	0.77	–	–	–	–
		ssGBLUP	–	_a0.78^a	_a0.78^a	_a0.77^b	_a0.77^b
		ssFGBLUP	–	–	_a0.79^a	_ab0.77^b	_a0.77^b
		ssWGBLUP_bin1	–	_b0.76^a	_b0.76^a	_bc0.76^a	_b0.76^a
		ssWGBLUP_bin30	–	_b0.76^b	_b0.77^a	_ab0.77^a	_ab0.76^ab
		ssWGBLUP_bin100	–	_b0.76^a	_b0.76^ab	_c0.75^bc	_c0.74^c
Protein	Genotyped	PBLUP	0.81	–	–	–	–
		ssGBLUP	–	_a0.76^a	_a0.77^a	_a0.76^a	_a0.76^a
		ssFGBLUP	–	–	_a0.78^a	_a0.75^b	_a0.76^b
		ssWGBLUP_bin1	–	_b0.73^a	_b0.73^a	_b0.72^a	_b0.72^a
		ssWGBLUP_bin30	–	_b0.73^a	_b0.74^a	_b0.72^a	_b0.73^a
		ssWGBLUP_bin100	–	_b0.74^a	_b0.73^ab	_b0.72^ab	_b0.72^b
	Non-genotyped	PBLUP	0.80	–	–	–	–
		ssGBLUP	–	_a0.77^a	_ab0.77^a	_a0.77^a	_a0.77^a
		ssFGBLUP	–	–	_a0.78^a	_ab0.76^b	_a0.77^ab
		ssWGBLUP_bin1	–	_a0.77^ab	_ab0.77^a	_ab0.76^ab	_ab0.76^b
		ssWGBLUP_bin30	–	_a0.76^a	_b0.76^a	_ab0.76^a	_a0.76^a
		ssWGBLUP_bin100	–	_a0.77^a	_b0.76^ab	_b0.75^bc	_b0.75^c

ssGBLUP models = a single-step genomic BLUP (ssGBLUP) model, a featured ssGBLUP (ssFGBLUP) model, and weighted ssGBLUP models with a region size of 1 (ssWGBLUP_bin1), 30 (ssWGBLUP_bin30), or 100 SNPs (ssWGBLUP_bin100)
SNP sets = SNPs on the 54K-SNP chip (54K), SNPs on the 54K-SNP chip together with sequencing SNPs selected by Denmark-Finland-Sweden (54K + DFS), SNPs on the 54K-SNP chip together with sequencing SNPs selected by France (54K + FRA), and SNPs on the 54K-SNP chip together with both sets of selected sequencing SNPs (54K + DFS + FRA)
_a,b,csubscript letters to the left are for comparisons among models using the same SNP set, and superscript letters to the right are for comparisons among SNP sets using the same model. Regression coefficients with no common letter differ significantly (P < 0.05)

Back to article page

ISSN: 1297-9686

Contact us

Submission enquiries: gse@inrae.fr
General enquiries: info@biomedcentral.com