Premium
Application of an Optimization Model to Multi‐Trait Selection Programs
Author(s) -
Johnson Blaine,
Gardner C. O.,
Wrede K. C.
Publication year - 1988
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/cropsci1988.0011183x002800050001x
Subject(s) - trait , selection (genetic algorithm) , ranking (information retrieval) , rank (graph theory) , flexibility (engineering) , biology , microbiology and biotechnology , statistics , breeding program , index selection , mathematics , computer science , machine learning , agronomy , programming language , combinatorics , cultivar
Simultaneous multi‐trait selection involves making many separate decisions which ultimately reflect the total value of each genotype being considered. A linear index is frequently used to simplify these decisions. However, use of traditional selection indices has required knowledge of the relative genetic merit of each trait. The objective of this investigation was to devise a method which applied breeders could utilize in determining appropriate relative values for the genetic merits. A procedure was developed whereby approximate values could be derived from a ranking of genotypes. The procedure requires a subjective ranking made by the breeder based on simultaneous consideration of the traits of interest. The utility of the procedure in applied breeding situations was tested. The technique was used to determine the relative merit of each of three traits, grain yield, moisture content of grain at harvest, and percent standing plants at harvest, in six sets of selected and ranked S 2 and S 3 families of maize. The selections and rankings were subjectively made by two experienced maize breeders. The two breeders were unable to numerically express acceptable relative merits, or relative weights, for the three traits prior to the solution of the problem. They were able to state general selection goals and to subjectively rank the families accordingly. The two breeders had basic philosophical differences in selection goals and these differences were detected by the weights derived from the breeders' respective rankings. The method provides flexibility which is lacking in other procedures and has potential utility in applied breeding programs as a management and operational tool.