Plant & Animal Genome V Conference
Town & Country Hotel, San Diego, CA, January 12-16, 1997.
PAG-V: W34 - THE THEORY AND PRACTICE OF MARKER-AIDED-SELECTION IN CONIFEROUS FOREST TREES
W34
THE THEORY AND PRACTICE OF MARKER-AIDED-SELECTION IN CONIFEROUS FOREST TREES
Carson, S.D., M.J. CARSON, P.L. Wilcox, T.E. Richardson
New Zealand Forest Research Institute, Private Bag 3020, Rotorua, New Zealand
The application of marker-aided-selection (MAS) has the potential to increase genetic gain in breeding programmes of coniferous forest trees above that realised from conventional population improvement methods. MAS will be particularly useful for high heritability traits that are expensive to measure on large numbers of trees (eg. wood property traits). While markers should provide more efficient selection of progenies for advancing breeding programmes, there will be delays in realising gain because of the long generation time. Gain from MAS will be captured more quickly in production populations by deploying selected juvenile clones in operational plantings produced from control-pollinated seed and vegetative multiplication. Gain from this strategy would be optimised if QTL detection experiments are carried out with offspring of the best parents (possible in forest trees because of the high genetic variance among offspring of any two parents). Because economically important traits in forest trees are continuously distributed with little suggestion of major gene action, the greatest difficulty in implementing MAS appears to be detecting biologically real QTL and obtaining a true estimate of the amount of variation explained. Simulation studies suggest that with the population sizes commonly used to locate marker-trait associations the size of QTL effects are often over-estimated, making additional verification experiments to confirm the existence and estimate the size of the effect of QTL essential. In order to achieve effective MAS among progeny of the top 10-20 highest ranked seed orchard parents, a small number of large QTL detection experiments with unrelated parents is suggested in order to detect most of the important QTL loci in the production population. For parents not in large QTL detection experiments, linkage phase would have to be determined by comparison of marker phenotypes with field trial data (since linkage equilibrium between loci must initially be assumed), but much smaller field experiments would be required for previously identified QTL. A SSR framework map composed of markers that segregate in most parents could reduce resources required to achieve effective MAS in production populations.