Table 1.

Average imputation accuracy from the bovine 777 K SNP chip to whole-genome sequence on chromosome 1 and 29

	Chromosome 1				Chromosome 29
	HOL20	MIX80	HOL80	MIX140	HOL20	MIX80	HOL80	MIX140
Sequence (n)	1,912,451	1,912,451	1,91,2451	1,912,451	670,773	670,773	670,773	670,773
777 K chip (n)	41,868	41,868	41,868	41,868	13,556	13,556	13,556	13,556
No variation in reference set (n)1	1,178,683	710,480	894,633	616,808	385,137	221,137	283,516	187,904
No variation observed in validation set (n)1	−2	468,203	284,050	561,875	−2	164,000	101,621	197,233
No variation imputed in validation set (n)1	19,484	1,005	1,077	649	19,284	4,139	4,267	3,681
Obtained overall imputation accuracy (n)	672,416	690,895	690,823	691,251	252,796	267,941	267,813	268,399
average overall imputation accuracy (r)	0.70	0.83	0.88	0.89	0.59	0.74	0.80	0.82
standard deviation of r	0.32	0.27	0.25	0.24	0.37	0.32	0.29	0.28

¹No variation was present in the genotype dosages of at least one of the 5 corresponding cross-validation sets, therefore the imputation accuracy (correlation) could not be computed.

²In scenario HOL20 the reference sets were the same as the validation sets, therefore all variants without variation in at least one cross-validation reference set are the same as the variants without variation in observed genotypes of the validation sets.