TABLE 7.

Genomic prediction accuracy using 11 different models for stripe rust resistance at four locations, and against five isolates at seedling stage using 90K SNP array and genotyping-by-sequencing (GBS) platform.

Markers		NWS.15	ISB.15	NWS.16	ISB.16	Pst.571242	Pst.571262	Pst.140202	Pst.571243
90K	BayesA	0.506 (0.052)	0.405 (0.053)	0.482 (0.07)	0.506 (0.038)	0.279 (0.113)	0.219 (0.074)	0.11 (0.05)	0.426 (0.059)
	BayesB	0.489 (0.055)	0.407 (0.05)	0.486 (0.07)	0.51 (0.044)	0.286 (0.111)	0.228 (0.079)	0.101 (0.051)	0.414 (0.062)
	BayesC	0.491 (0.061)	0.413 (0.052)	0.477 (0.077)	0.5 (0.044)	0.29 (0.104)	0.236 (0.076)	0.107 (0.055)	0.434 (0.061)
	BRR	0.488 (0.055)	0.395 (0.056)	0.502 (0.061)	0.498 (0.044)	0.264 (0.102)	0.236 (0.079)	0.139 (0.041)	0.413 (0.06)
	BL	0.511 (0.053)	0.392 (0.047)	0.47 (0.078)	0.493 (0.043)	0.275 (0.101)	0.24 (0.073)	0.102 (0.05)	0.42 (0.059)
	GBLUP	0.468 (0.058)	0.393 (0.048)	0.472 (0.069)	0.494 (0.044)	0.259 (0.103)	0.241 (0.077)	0.106 (0.038)	0.396 (0.059)
	RKHS	0.354 (0.054)	0.235 (0.054)	0.39 (0.081)	0.386 (0.062)	0.205 (0.094)	0.221 (0.064)	0.061 (0.049)	0.226 (0.084)
	EN	0.48 (0.062)	0.413 (0.06)	0.482 (0.069)	0.491 (0.043)	0.227 (0.104)	0.214 (0.085)	0.117 (0.033)	0.402 (0.056)
	RVM	0.505 (0.067)	0.396 (0.072)	0.486 (0.069)	0.466 (0.049)	0.245 (0.108)	0.212 (0.06)	0.131 (0.043)	0.34 (0.054)
	GP	0.476 (0.058)	0.392 (0.06)	0.5 (0.058)	0.498 (0.048)	0.258 (0.107)	0.25 (0.076)	0.088 (0.038)	0.408 (0.057)
	RRBLUP	0.481 (0.057)	0.406 (0.052)	0.486 (0.069)	0.501 (0.042)	0.228 (0.099)	0.222 (0.078)	0.139 (0.035)	0.405 (0.058)
GBS	BayesA	0.449 (0.108)	0.442 (0.077)	0.391 (0.061)	0.399 (0.089)	0.168 (0.047)	0.102 (0.07)	0.079 (0.053)	0.23 (0.051)
	BayesB	0.421 (0.118)	0.432 (0.083)	0.386 (0.06)	0.405 (0.086)	0.146 (0.046)	0.117 (0.067)	0.067 (0.056)	0.227 (0.048)
	BayesC	0.421 (0.114)	0.412 (0.084)	0.395 (0.061)	0.396 (0.084)	0.157 (0.053)	0.122 (0.07)	0.097 (0.053)	0.236 (0.044)
	BRR	0.407 (0.114)	0.426 (0.079)	0.398 (0.061)	0.428 (0.084)	0.146 (0.05)	0.106 (0.067)	0.062 (0.059)	0.229 (0.048)
	BL	0.428 (0.111)	0.399 (0.087)	0.38 (0.062)	0.391 (0.093)	0.15 (0.05)	0.109 (0.071)	0.054 (0.056)	0.241 (0.048)
	GBLUP	0.417 (0.111)	0.435 (0.081)	0.366 (0.069)	0.388 (0.086)	0.151 (0.051)	0.117 (0.069)	0.013 (0.054)	0.225 (0.05)
	RKHS	0.403 (0.115)	0.424 (0.082)	0.36 (0.064)	0.407 (0.09)	0.104 (0.043)	0.107 (0.07)	0.039 (0.052)	0.228 (0.052)
	EN	0.26 (0.086)	0.33 (0.109)	0.17 (0.104)	0.379 (0.066)	0.256 (0.049)	0.002 (0.087)	−0.01(0.07)	0.084 (0.065)
	RVM	0.486 (0.106)	0.381 (0.08)	0.346 (0.072)	0.439 (0.084)	0.06 (0.04)	0.038 (0.071)	0.171 (0.06)	0.224 (0.083)
	GP	0.413 (0.114)	0.427 (0.087)	0.372 (0.069)	0.466 (0.084)	0.139 (0.049)	0.133 (0.071)	0.048 (0.066)	0.246 (0.056)
	RRBLUP	0.398 (0.109)	0.421 (0.082)	0.388 (0.061)	0.385 (0.087)	0.079 (0.056)	0.104 (0.07)	0.001 (0.053)	0.213 (0.055)

Genomic prediction models: BayesA, BayesB, and BayesC. BRR, Bayesian ridge regression; BL, Bayesian least absolute shrinkage and selector operator; GBLUP, genomic best linear unbiased prediction; RKHS, reproducing kernel Hilbert spaces regression; EN, elastic net; RVM, relevance vector machine; GP, Gaussian processor; rrBLUP, ridge regression best linear unbiased prediction. The values in the parentheses are SDs of the prediction accuracies.