Assessing accuracy of imputation using different SNP panel densities in a multi-breed sheep population

Table 3 Imputation scenarios from 5K to 50K (50K original) using two types of reference population

Scenario	Number of reference animals	Number of imputed animals	Description of reference animals^d	Imputed group breed^b
25_5K50K	15,443^a and 4564^b	218	All breeds/Romney	Romney 100 %
26_5K50K	15,443^a and 4326^b	142	All breeds/Romney	Romney < 65 %
27_5K50K	4256	1000^c	Romney	Romney
28_5K50K	15,443^a and 2324^b	250	All breeds/Coopworth	Coopworth 100 %
29_5K50K	15,443^a and 2279^b	250	All breeds/Coopworth	Coopworth < 70 %
30_5K50K	15,443^a and 640^b	250	All breeds/Perendale	Perendale > 95 %
31_5K50K	15,443^a and 138^b	172	All breeds/Composites	Composites > 50 % < 95 %

^aFixed reference population that included 15,443 animals from all breeds with genotyped animals
^bWithin-breed/group reference population: some groups contained a small number of genotyped animals
^c1000 animals defined as the imputed set to optimize the calculation of the r² imputation accuracy per SNP
^dTwo types of reference population were used: (1) a fixed reference population that included a large number of animals from all breeds and (2) a within-group reference population

ISSN: 1297-9686