Genetic gain and diversity of seed orchards' crops are determined by the number of parents, their breeding values and relatedness, within-orchard pollination efficiency, and level of pollen contamination. These parameters can be manipulated at establishment by varying clonal representation (e.g., linear deployment), during orchard development by genetic thinning, or by selective harvesting. Since clonal fecundities are known to vary both within and among years, then each seed crop has a unique genetic composition and, therefore, crops should be treated on a yearly basis. Here we present an optimization protocol that maximizes crop's genetic gain at any desired genetic diversity through the selection of a subset of the crop that meets both parameters. The genetic gain is maximized within the biological limit set by each clone's seed-cone production and effective population size is used as a proxy to genetic diversity whereby any relationship among clones is considered. The optimization was illustrated using 3 years' reproductive output data from a first-generation western larch seed orchard and was tested under various scenarios including actual male and female reproductive output and male reproductive output assumed to be either equal to that of female or a function of clonal representation. Furthermore, various levels of co-ancestry were assigned to the orchard's clones in supplementary simulations. Following the optimization, all solutions were effective in creating custom seedlots with different gain and diversity levels and provided the means to estimate the genetic properties of composite seedlots encompassing the remaining "unused" seed from a number of years.

No abstract

Strong convergence, M-estimators, epi-convergence, stochastic optimisation,

No abstract