Accuracy of genotype imputation based on random and selected reference sets in purebred and crossbred sheep populations and its effect on accuracy of genomic prediction

Table 1 Accuracy of genomic prediction based on observed genotypes (50 or 12k) and imputed 50k genotypes with different accuracies for post-weaning weight (PWW), post-weaning eye muscle depth (PW_EMD) and yearling greasy fleece weight (YGFW) in Merino sheep

Genotypes	Imputation accuracy	PWW	PW_EMD	YGFW
Observed 50k	NA^a	0.446	0.219	0.585
Imputed 50k − 1^b	0.88–0.99	0.443	0.219	0.584
Imputed 50k − 2(1)^c	0.73–0.96	0.428	0.217	0.583
Imputed 50k − 2(2)^d	0.73–0.96	0.430	0.215	0.582
Imputed 50k − 3^e	0.62–0.86	0.394	0.184	0.572
Imputed 50k − 4^f	0.48–0.60	0.381	0.171	0.534
Observed 12k	NA	0.412	0.205	0.552

^aNot applicable
^bImputed from 3000 purebred Merino
^cImputed from 3000 crossbred Merino
^dImputed from 1000 crossbred Merino and edited for individual SNPs with low imputation accuracy (r < 0.7)
^eImputed from 1000 mixed crossbred Merino
^fImputed from 367 non-Merino purebreds

ISSN: 1297-9686