Premium
Breeding Line Selection Based on Multiple Traits
Author(s) -
Yan Weikai,
FrégeauReid Judith
Publication year - 2008
Publication title -
crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.76
H-Index - 147
eISSN - 1435-0653
pISSN - 0011-183X
DOI - 10.2135/cropsci2007.05.0254
Subject(s) - biplot , selection (genetic algorithm) , trait , biology , avena , microbiology and biotechnology , culling , breeding program , quantitative trait locus , genomic selection , index selection , genotype , genetics , computer science , agronomy , machine learning , ecology , herd , gene , cultivar , single nucleotide polymorphism , programming language
Breeding line selection, either for potential varieties or for useful parents, must be based on multiple breeding objectives (or traits). Varieties cannot have any major defects, while parents must have outstanding levels in at least one trait. Due to undesirable associations among breeding objectives, it is difficult to accomplish both tasks (variety selection and parent selection) through a single selection strategy. Additional complication results when a program is breeding for different end‐uses such that both high and low levels of a trait are desirable. The first purpose of this paper was to propose a comprehensive multitrait selection procedure that coherently combines independent selection, independent culling, and index selection so that all the aspects in breeding line selection are taken into consideration. A dataset of 150 oat ( Avena sativa L.) breeding lines with values evaluated for four quality traits (groat, oil, protein, and beta‐glucan concentrations) was used for illustration. A genotype by trait biplot is a useful tool for exploring multiple trait data and can aid in multitrait selection because it graphically displays the trait associations across, and the trait profiles of, the genotypes. Procedures are outlined to avoid possible misinterpretation of such a biplot when the biplot does not fully display the patterns.