Alternative methods improve the accuracy of genomic prediction using information from a causal point mutation in a dairy sheep model

Table 5 Description of the different genetic evaluation models based on a single-step approach and using information about the SOCS2 gene or not

Approach	Model	Use of SOCS2 data	Information used in the relationship matrix
Approach	Model	Use of SOCS2 data	Pedigree	54K SNPs	SOCS2 SNP
Single-trait	Pedigree-based BLUP	No	Yes	No	No
	ssGBLUP	No	Yes	Yes	No
	ssGBLUP_SOCS2^a	Yes	Yes	Yes	Yes
	WssGBLUP_{(m, n)}^b	No	Yes	Yes	No
	WssGBLUP_{SOCS2 (m, n)}^b	Yes	Yes	Yes	Yes
Multiple-trait	Pedigree-based Gene Content	Yes (as a trait)	Yes	No	No

Abbreviations: GBLUP Genomic Best Linear Unbiaised Prediction, ss single-step, W Weighted
^aThe term SOCS2 here means that the SOCS2 SNP has been added to the 54K SNPs of the chip
^bFour approaches to the WssGBLUP were computed (m = classical, mean, maximum or sum). The classical WssGBLUP approach (m = classical) gives a different weight for each marker of the chip. In alternative approaches, the chip is decomposed into non-overlapping windows of n markers (we tested n = 2, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 100, and 200) and within these windows, all markers are assigned the same weight: the mean weight of the n SNPs (m = mean), the maximum weight of the n SNPs (m = maximum), and the sum of the n SNPs weights (m = sum)

ISSN: 1471-2164