Financial feasibility of marker-aided selection in Douglas-fir

The land area required for a marker-aided selection (MAS) program to break-even (i.e., have equal costs and benefits) was estimated using computer simulation for coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) in the Pacific Northwestern United States. We compared the selection efficiency obtained when using an index that included the phenotype and marker score with that obtained using only the phenotype. It was assumed that MAS was restricted to within-family selection, that the rotation age was 50 years, and that growth rate (h(2) = 0.25), tree form (h(2) = 0.25), and (or) wood density (h(2) = 0.45) were the objects of improvement. Several population quantitative trait loci (QTL) models, selection population sizes, and interest rates were considered. When large selection population sizes were employed (500 trees per family) MAS gave considerable increases in efficiency of within-family selection; however, results showed that the combination of small selection population sizes (100 trees per family) and many QTL of moderate effect could lead to losses in gain from MAS compared with phenotypic selection. For many reasonable selection scenarios and the simplified assumptions in our model, the land base required for breeding programs to break-even is smaller or near to the limit of those in place under operational breeding programs in the region. Considerably more research is needed to reasonably predict whether MAS would be cost-effective in practice. However, before some of the basic research needed to implement MAS can be done, organizations need to establish large blocks of full-sib families to allow for QTL identification.