Genomic selection can replace phenotypic selection in early generation wheat breeding
| Autor: | Borrenpohl, Daniel |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: | |
| Druh dokumentu: | Text |
| Popis: | Genomic selection (GS) enables plant breeders to estimate the value of unphenotyped lines using phenotypic and genotypic data from phenotyped lines. Plant breeders are investigating if GS can replace phenotypic selection (PS) in Stage-1 trials (also known as preliminary trials) where lines are phenotyped with minimal replication in one or few locations. We genotyped 1769 lines from the OSU winter wheat breeding program and used historical data from 2013-2018 to test various GS methods and compare them to PS. Our objectives were to: 1) test the prediction accuracy of GS as compared to PS; 2) test the impact of Stage-1 phenotypic data on GS prediction accuracy; and 3) to compare selections from PS and GS in Stage-1 trials. GS prediction accuracy for grain yield and fusarium head blight (FHB) index was comparable to PS regardless of Stage-1 phenotypic data used. GS methods using no Stage-1 phenotypic data (NST1) had marginally lower rank equivalence and selection coincidence for grain yield and FHB index compared to GS methods using Stage-1 phenotypic data and PS. Economic analysis suggests that using NST1 for Stage-1 selections could triple Stage-1 trial size compared to PS using equivalent resources. Given a fixed number of lines are selected for Stage-2 trials, higher selection intensity is observed for NST1 than PS. As a result, higher estimates of indirect gain from selection were observed for NST1 than PS for grain yield and FHB index. Our results suggest that GS can potentially replace PS in making selections from Stage-1 trials. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|