Effects of random and non-random errors on phenotypic selection in autotetraploids.

摘要: A theoretical investigation was made to ascertain the effects of random and non-random deviations, called errors, phenotypic from genotypic values on population means response recurrent selection. The study motivated as a selection experiment for disease resistance where there either variability in inoculation or environment (the errors) above below optimum rate genetic differences are maximized errors). limited genetics at diallelic locus (alleles B b) an autotetraploid mating equilibrium. measured covariance compared exact which without errors phenotype. were modeled by assuming that given percentage (α) uniformly distributed among five possible genotype classes independent their true genotypes. This model analyzed numerically with frequency allele (p) ranging 0.0 1.0 assumed ofα=0.1 0.5 following six types genic action allele: additive, monoplex dominance, partial duplex recessive. effect error consistently reduce type locus. mean upward bias when p low downward approached unity. In inoculations altered phenotypes systematically includingα susceptible genotypes into one more other (a positive shift). With inoculations, some resistant classed non-resistant negative found numerical analysis same andα's investigated error. shift p, always reduced, but increasing exceeded all except additive. greatly improved three action: upwards shift, decreased. relatively slight decrease occurred shift. indicated check varieties commonly used monitor pressures screening programs very responsive shifts, unresponsive shifts. interaction pressure, action, class models suggested explanation lack observed breeding programs.