MODELING AND OPTIMIZING OF SEQUENTIAL SELECTION SCHEMES - A POULTRY BREEDING APPLICATION

A sequential selection scheme, where candidates are ranked using a mul tiple trait BLUP selection index, was modelled deterministically. This model accounts for overlapping generations and for the reduction of g enetic variances under selection, in order to predict the asymptotic g enetic gain. Sires and darns are selected among the pairs already crea ted whose progeny have maximum expected average genetic merit. This pr ocedure allows for an optimal use of the available information when th e pairs are selected. Effects of selection on the mean and variance of the traits measured on selected animals are accounted for using the T allis formulae, while a matrix formula is used in order to simultaneou sly derive genetic lags and gains. The evolution of inbreeding rate wa s not modelled. Numerical applications were related to a turkey breedi ng plan. The impact of the relative weight given to growth (male and f emale body weight, measured at 12 and 16 weeks) and reproduction trait s (three partial egg number records) on the expected genetic gains was investigated. Influence of demographic parameters was also studied. D ifferent selection strategies were compared. When the selection object ive is mainly to improve laying ability, it is more relevant to increa se the amount of information on laying performance: and to apply selec tion of best mated pairs, rather than to reduce generation intervals b y only using the youngest sires. This modelling can be viewed as a use ful tool, in order to foresee the consequences of any change in the br eeding plan for the long-term genetic gain.