Maximising genetic response in tree breeding with constraints on group coancestry

This study compares 4 methods to maximize response to selection while minimizing the inbreeding that accumulated over 6 generations of selection. A selection algorithm is described which maximizes the mean estimated breeding value of the selected trees while constraining the average coancestry among them and hence the status effective number of the population. This algorithm resulted in increased selection response, at a lower, target average coancestry level, compared to truncation selection. However the target average coancestry level for each generation is arbitrarily chosen. A variation of the algorithm is described which attempts to maximize response over 12 generations allowing for the decline in genetic variance caused by inbreeding. This leads to faster genetic gain and higher coancestry than typical tree breeding programs. Once the selection has been carried out, a mate allocation algorithm arranges matings which give each parent the number of offspring required by the selection algorithm and which minimizes inbreeding in the offspring per mating. This delays the appearance of inbreeding by at least one generation but does not alter the long term rate of inbreeding. The combination of the selection and mating algorithms lead to breeding programs requiring less controlled pollinations than truncation selection with random mating and hence reduce the cost of the breeding program.