Premium
Selection Indices for the Improvement of Opaque‐2 Maize 1
Author(s) -
St. Martin S. K.,
Loesch P. J.,
DemopulosRodriguez J. T.,
Wiser W. J.
Publication year - 1982
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/cropsci1982.0011183x002200030010x
Subject(s) - selection (genetic algorithm) , biology , trait , kernel (algebra) , statistics , culling , mathematics , microbiology and biotechnology , agronomy , machine learning , computer science , zoology , herd , combinatorics , programming language
The need to improve agronomic and protein and kernel quality characters in opaque‐2 maize ( Zea mays L.) mandates consideration of multiple traits in selection. Data from genetic studies in an opaque‐2 maize synthetic were used to calculate selection indices for use in recurrent selection. Grain yield was considered the only trait having explicit economic value. Restricted indices were used to achieve satisfactory gains in kernel hardness, moisture content, and protein content and quality. The most important traits with respect to achieving the specified selection goals were yield, moisture content, kernel hardness, lysine content, kernel weight, and the degree of light transmission by kernels. Indices incorporating five traits were more efficient in that the addition of a sixth trait contributed very little to expected gains. The most effective combinations of traits for a recurrent selection scheme involving S 1 testing were also the most effective for S 2 testing and full‐sib testing. The degree to which gains predicted for an index were dependent on accurate estimates of the population parameters was considered. Indices having weights that implied reliance on indirect rather than direct selection response were not recommended. Substitution of single‐plant data for data from replicated progeny tests was considered in connection with each breeding scheme. The possibility of increasing the effective selection intensity, given that resources for replicated testing were fixed, by culling individual plants was investigated, but the use of individual plant data did not increase expected gain. A measure of subjective judgment on the part of the breeder is probably required for the successful application of selection indices to complex problems.