A simulation program for the design of broiler nucleus breeding strategies

Thumbnail Image



Mehrabani-Yeganeh, Hassan

Journal Title

Journal ISSN

Volume Title


University of Guelph


A computer simulation program (BBSIM) was developed that can be used to explore breeding strategies for poultry nucleus populations. The program can create multiple traits with different genetic controls (polygenic, finite locus model, categorical, and mixed inheritance). Several stages in the life of an animal can be defined and traits can be expressed or not expressed at any stage. A dynamic best linear unbiased prediction (BLUP) evaluation that can handle single and multiple trait situations is provided. The coefficient of inbreeding is calculated for each individual. Culling and selection can be based on BLUP estimated breeding value (EBV), an index of EBV, or phenotype. Matings can be random or assortative. The program was used to study the effect of less stringent convergence criterion (CC) in calculation of EBV on genetic response. The sum of squares of differences between consecutive solutions divided by sum of squares of new solutions was used as the criterion for terminating the iteration process. The average breeding values of the top 2% of individuals increased as the system moved towards the most stringent CC. The gain in average breeding value, however, was very small, beyond a CC of 10\sp−3 and could not justify the extra number of iterations and CPU time that were needed to reach 10\sp−7. The program was used to examine the effect of including different amounts of pedigree and historical data on genetic response under single trait BLUP. Using the last two generations of data and pedigree, under a broiler breeding family structure, produced genetic response that was not significantly different from full data and pedigree (10 generations) The amount of CPU time, however, was reduced drastically. The results are of particular importance in studies using simulation. Whether inbreeding coefficients should be considered or ignored in the construction of A\sp−1 was another question that was addressed using the BBSIM program. With rapid rates of inbreeding, under population structures representative of broiler breeding, pairwise comparisons of average genetic responses found no significant differences in response when inbreeding was accounted for in A\sp−1. In a two-trait two-stage selection scheme mimicking selection on body weight and egg number, different culling procedures were applied at stage one. Culling was either on BLUP EBV or phenotype and information of culled animals were either carried forward to the second stage or not carried forward. Culling based on BLUP produced greater response, except when the negative genetic correlation between the two traits was high. When information of the culled individuals was not carried to the next stage, genetic response decreased significantly. It was concluded that there are significant advantages to capturing and utilising information from the first stage selection.



Computer simulation program, Breeding strategies, Poultry nucleus populations, Multiple traits, Different genetic controls