Fixed-length haplotypes can improve genomic prediction accuracy in an admixed dairy cattle population

Table 3 Computation time and number of random covariates in haplotype and SNP BayesA models

Trait	Freq^a	Number of random covariates					Computation time (h)^b
Trait	Freq^a	125 kb	250 kb	500 kb	1 Mb	2 Mb	125 kb	250 kb	500 kb	1 Mb	2 Mb
Milk fat yield	SNP	37,226	37,226	37,226	37,226	37,226	13.1	13.1	13.1	13.1	13.1
	1%	56,590	64,724	70,380	56,534	32,520	22.8	23.5	24.7	20.0	11.3
	2.5%	51,889	53,482	47,378	29,343	13,460	21.3	19.7	16.8	10.4	4.8
	5%	46,283	41,737	28,324	12,291	3977	19.6	15.5	10.4	4.5	1.5
	10%	37,848	27,656	12,790	3255	646	15.2	10.8	5.0	1.4	0.3
Liveweight	SNP	37,356	37,356	37,356	37,356	37,356	6.6	6.6	6.6	6.6	6.6
	1%	56,595	64,634	70,218	56,164	32,117	11.0	13.1	13.3	9.9	5.7
	2.5%	51,839	53,204	46,797	28,756	13,050	10.2	9.6	9.2	5.2	2.4
	5%	46,163	41,467	28,040	12,198	4027	9.2	7.7	5.2	2.3	0.8
	10%	37,775	27,604	12,882	3354	707	7.8	5.4	2.6	0.7	0.2
Somatic cell score	SNP	37,229	37,229	37,229	37,229	37,229	13.0	13.0	13.0	13.0	13.0
	1%	56,630	64,730	70,375	56,521	32,516	21.4	24.4	27.2	19.7	11.1
	2.5%	51,934	53,488	47,385	29,348	13,464	23.1	20.8	16.7	10.9	4.7
	5%	46,326	41,746	28,329	12,296	3977	18.3	15.4	10.2	4.5	1.5
	10%	37,898	27,663	12,793	3254	645	15.1	10.7	5.0	1.3	0.3

^aFrequency threshold for removing rare haplotype alleles. SNP refers to fitting covariates for SNPs rather than haplotype alleles
^bComputation time for running the analysis on the training set containing all breeds with a chain length of 41,000

ISSN: 1297-9686