Maximizing efficiency in sunflower breeding through historical data optimization

Fernández-González, Javier; Haquin, Bertrand; Combes, Eliette; Bernard, Karine; Allard, Alix; Isidro y Sánchez, Julio

doi:10.1186/s13007-024-01151-0

Plant Methods

Table 2 Overview of the average performance of the optimization methods in all traits and scenarios measured as the percentage of gain in area under the curve (AUC) relative to random sampling

From: Maximizing efficiency in sunflower breeding through historical data optimization

Optimization type	Optimization methods	Test set year 6						Test set year 7							Global average	Variance	Trait
Optimization type	Optimization methods	CS year 5	CS year 4-5	CS year 3–5	CS year 2–5	CS year 1–5	Test set average	CS year 6	CS year 5–6	CS year 4-6	CS year 3–6	CS year 2–6	CS year 1–6	Test set average	Global average	Variance	Trait
Genetic-based	Avg_GRM_self	− 1.17	2.89	1.11	− 0.94	− 0.11	0.36	2.25	3.46	1.51	0.13	− 0.5	0.47	1.22	0.83	2.41	YLD
	Avg_GRM_MinMax	0.37	2.35	0.78	0.12	0.64	0.85	0.54	3.64	1.6	− 0.44	− 0.58	0.8	0.93	0.89	1.52
	PCA_CDmean	0.25	1.94	0.51	1.43	1.38	1.1	0.27	2.54	1.07	− 1.96	− 0.93	− 1.43	− 0.07	0.46	2.01
Mixed	PLS_CDmean	0.6	1.86	2.1	0.01	1.8	1.27	− 0.06	3.41	1.21	− 1.63	− 1.33	− 0.25	0.23	0.7	2.36
Mixed	Tails_GEGVs	0.3	4.83	7.66	4.55	7.35	4.94	2.22	3.02	1.48	− 1.35	1.55	0.66	1.26	2.93	8.31
Phenotypic-based	Tails	0.53	5.98	3.72	2.38	4.07	3.34	4.93	6.38	5.97	3.99	3.93	3.83	4.84	4.16	2.88
Phenotypic-based	Scenario average	0.15	3.31	2.65	1.26	2.52	1.98	1.69	3.74	2.14	− 0.21	0.36	0.68	1.4	1.66	1.76
	Test set average			1.98						1.4					1.66	0.09
Genetic-based	Avg_GRM_self	1.18	1.85	2.06	0.41	− 0.34	1.03	0.05	− 0.38	1.53	0.53	2.31	1.48	0.92	0.97	0.93	GM
	Avg_GRM_MinMax	1.52	2.18	1.43	0.55	0.65	1.27	1.08	− 0.05	0.98	1.18	2.17	1.08	1.07	1.16	0.44
	PCA_CDmean	− 0.74	0.43	0.48	− 0.42	− 0.63	− 0.18	− 0.57	0.52	0.37	− 0.13	1.62	0.24	0.34	0.11	0.48
Mixed	PLS_CDmean	0.03	− 0.06	− 0.33	− 0.43	− 1.27	− 0.41	− 1.28	− 1.37	− 0.25	− 0.74	0.18	0.98	− 0.41	− 0.41	0.52
Mixed	Tails_GEGVs	− 3.82	0.2	− 1.19	0.71	1.09	− 0.6	2.49	1.69	− 2.22	− 3.16	− 5.06	− 4.58	− 1.81	− 1.26	7.09
Phenotypic-based	Tails	− 8.5	− 0.12	− 1.31	0.06	0.92	− 1.79	3.8	2.11	1.54	1.73	0.89	0.43	1.75	0.14	9.98
Phenotypic-based	Scenario Average	− 1.72	0.75	0.19	0.15	0.07	− 0.11	0.93	0.42	0.32	− 0.1	0.35	− 0.06	0.31	0.12	0.47
	Test set Average			− 0.11							0.31				0.12	0.05
Genetic-based	Avg_GRM_self	3.11	1.59	2.05	0.94	1.41	1.82	0.72	− 0.61	− 0.05	− 0.49	0.46	− 0.37	− 0.06	0.8	1.38	OIL
	Avg_GRM_MinMax	2.08	1.04	1.36	1.04	1.75	1.45	1.77	− 0.7	0.26	− 0.42	− 0.02	0.05	0.16	0.75	0.91
	PCA_CDmean	0.06	1.49	0.91	0.45	0.84	0.75	1.51	0.71	0.48	0.26	0.41	− 0.49	0.48	0.6	0.35
Mixed	PLS_CDmean	0.34	0.27	0.89	0.49	0.35	0.47	3.01	1.08	− 0.72	− 0.63	− 0.09	− 0.34	0.38	0.42	1.06
Mixed	Tails_GEGVs	− 11.97	− 3.34	− 1.31	− 1.49	− 2.04	− 4.03	1.52	− 3.67	− 4.23	− 3.9	− 1.5	− 0.22	− 2.00	− 2.92	11.97
Phenotypic-based	Tails	− 8.1	0.49	− 0.67	− 1.39	0.47	− 1.84	2.78	− 0.08	0.05	1.5	0.62	0.46	0.89	− 0.35	7.77
Phenotypic-based	Scenario Average	− 2.41	0.26	0.54	0.01	0.46	− 0.23	1.89	− 0.54	− 0.7	− 0.61	− 0.02	− 0.15	− 0.02	− 0.12	1.1
	Test set Average			− 0.23							− 0.02				− 0.12	0.01

The optimization methods are classified based on the type of input they require, i.e. Genetic-based, Phenotypic-based, and Mixed. The “Scenario Average” rows display the average performance of each scenario (test set and candidate set years combination) across optimization methods. The “Test set Average” column provides the average performance within a given test set across candidate sets, while the “Test set Average” rows display the average performance across optimization methods. Additionally, the figure presents the global average performance across scenarios for each optimization method, along with its corresponding variance

Back to article page

ISSN: 1746-4811

Contact us

Submission enquiries: journalsubmissions@springernature.com