Reduction in accuracy of genomic prediction for ordered categorical data compared to continuous observations

Table 3 Spearman rank correlation among predicted ( $ĝ_{Con}, ĝ_{T}$ and $ĝ_{L}$ ) genotypic values

		$r {(ĝ_{Con}, ĝ_{T})}_{PV}$	$r {(ĝ_{Con}, ĝ_{L})}_{PV}$	$r {(ĝ_{T}, ĝ_{L})}_{PV}$	$r {(ĝ_{Con}, ĝ_{T})}_{MV}$	$r {(ĝ_{Con}, ĝ_{L})}_{MV}$	$r {(ĝ_{T}, ĝ_{L})}_{MV}$
h ²	c	n = 1000	n = 1000	n = 1000	n = 1000	n = 1000	n = 1000
0.10	2	0.26	0.21	0.61	0.29	0.26	0.60
	3	0.27	0.22	0.62	0.31	0.27	0.57
	4	0.32	0.26	0.52	0.34	0.29	0.48
0.25	2	0.34	0.31	0.74	0.33	0.29	0.73
	3	0.37	0.34	0.74	0.37	0.34	0.70
	4	0.40	0.33	0.64	0.40	0.29	0.56
0.50	2	0.54	0.45	0.66	0.50	0.42	0.66
	3	0.59	0.47	0.66	0.54	0.43	0.64
	4	0.61	0.51	0.64	0.55	0.46	0.61
h ²	c	n = 2250	n = 2250	n = 2250	n = 2250	n = 2250	n=2250
0.10	2	0.34	0.31	0.89	0.31	0.25	0.83
	3	0.40	0.35	0.88	0.34	0.28	0.82
	4	0.41	0.37	0.82	0.36	0.29	0.74
0.25	2	0.52	0.44	0.86	0.45	0.37	0.85
	3	0.54	0.47	0.84	0.48	0.41	0.82
	4	0.59	0.50	0.80	0.53	0.44	0.80
0.50	2	0.70	0.63	0.85	0.66	0.59	0.84
	3	0.72	0.66	0.87	0.68	0.62	0.84
	4	0.74	0.68	0.87	0.70	0.63	0.84

Results are for Angus (PV) and multibreed (MV) validation populations using Bayes C π linear (Con) model analysis of continuous phenotypes and Bayes C π linear (L) or threshold (T) model analysis of ordinal categorical phenotypes classified (c) as 2, 3 or 4 categories with heritabilities (h²) of 0.10, 0.25 or 0.50 in the Angus training (PT) population with (n) 1000 or 2250 observations.

ISSN: 1297-9686