BMC Genomics

Table 6 Predictive abilities of observed and predicted values of a 5 -fold-CV by year-location combination of models at first stage (M1, ⋯ , M9) for grain dry matter yield (Y), and repeatability ( R ) of the trait by location

From: The importance of phenotypic data analysis for genomic prediction - a case study comparing different spatial models in rye

Year	Loc	M1	M2	M3	M4	M5	M6	M7	M8	M9	R
2009	G-L1	0.469	0.473	0.462	0.481	0.448	0.455	0.474	0.481	0.478	0.376
2009	G-L2	0.271	0.272	0.279	0.280	0.282	0.288	0.282	0.270	0.269	0.177
2009	G-L3	0.347	0.344	0.351	0.350	0.345	0.339	0.350	0.355 ^§	0.355	0.264
2009	G-L4	0.595	0.593	0.597	0.602 ^§	0.592	0.594	0.602	0.592	0.598	0.440
2009	G-L5	0.495	0.514	0.506	0.505	0.519	0.527	–	0.514	0.529	0.303
2009	G-L6	0.393	0.398	0.357	0.372	0.359	0.360	0.369	0.372	0.378	0.077
2009	G-L7	0.596	0.594	0.586	0.599	0.578	0.565	0.591	0.584	0.577	0.299
2009	P-L1	0.127	0.118	0.132	0.138	0.116	0.114	0.138	0.174	0.167	0.225
2009	P-L2	0.301	0.306	0.303	0.310	0.307	0.309	0.310	0.323	0.323 ^§	0.338
2009	P-L3	-0.154	-0.165	-0.153	-0.154	-0.169	-0.172	-0.154	-0.158	-0.175	0.247
2009	P-L4	0.520	0.518	0.527	0.525	0.520	0.522	0.525	0.558	0.555	0.362
2010	G-L1	0.428	0.471	0.426	0.432	0.464	0.478	0.431	0.466	0.475	0.263
2010	G-L2	0.394	0.392	0.399	0.407	0.400	0.398	0.406	0.401	0.400	0.248
2010	G-L4	0.470	0.472	0.477 ^§	0.476	0.478	0.477	0.477	0.404	0.424	0.326
2010	G-L5	0.469	0.485	0.471	0.469	0.476	0.486	0.469	0.479	0.487	0.407
2010	G-L6	0.576	0.583	0.601	0.612	0.601	0.608	0.611	0.619	0.618	0.310
2010	G-L7	0.520	0.552	0.557	0.564	0.541	0.556	0.565	0.579	0.574	0.298
2010	G-L8	0.589	0.600	0.599	0.597	0.605	0.605	0.598	0.603	0.607	0.540
2010	P-L1	0.327	0.334	0.327	0.327	0.326	0.333	0.327	0.327	0.337	0.439
2010	P-L2	0.277	0.310	0.275	0.266	0.275	0.309	0.268	0.311	0.307	0.436
2010	P-L3	0.461	0.466	0.461	0.462	0.459	0.467	0.461	0.459	0.467	0.416
2010	P-L4	0.314	0.322	0.317	0.316	0.315	0.317	0.317	0.317	0.315	0.360
Counts		0	2	1	3	0	3	0	7	6
		0%	9%	5%	14%	0%	14%	0%	32%	27%

Boldfaced entries in the table indicate best model (fit) within location. Empty cells correspond to locations where the model did not converge. In italics, we report the models that converged but the Hessian matrix was not positive definite. § Better than second best model at forth decimal place

Back to article page

ISSN: 1471-2164

Contact us

Submission enquiries: bmcgenomics@biomedcentral.com
General enquiries: ORSupport@springernature.com