Genetic architecture of economic traits in Eucalyptus globulus

This thesis addresses several key issues related to the genetic architecture of traits relevant to the genetic improvement of Eucalypts globulus for pulpwood and solid wood production. It comprises three main studies, that address (i) the long-term effect of inbreeding, (ii) the genetic association between pulp-wood and solid wood selection traits and (iii) the importance of non-additive genetic effects. The long-term effect of inbreeding was studied using a 28 year-old field trial comprising outcrossed, selfed and open-pollinated (OP) families of Eucalyptus globulus and a commonly cooccurring species E. ovata. These species have a mixed mating system, where open-pollinated (OP) progenies are expected to include selfs and outcrosses. Inbreeding depression for early age growth due to selfing was high (age 4: 27% for E. globulus and 49% for E. ovata) but diminished with age, as size-dependent mortality resulted in the purging of smaller inbred progenies. Most mortality occurred between the ages 4 and 13 years, and size-dependent mortality resulted in a shift in inbreeding depression from growth to survival with age. OP progenies exhibited intermediate levels of inbreeding depression, but later-age survivors exhibited no evidence of inbreeding depression, consistent with the purging of selfs. General higher mortality rate in all cross-types of E. ovata compared with E. globulus was suggested to be due to climatic maladaptation, arising from the onset of drought conditions after 10 years of comparable growth, with the inbred progeny of both species appearing more susceptible. To further explore the genetics of growth and wood properties of E. globulus, two OP progeny trials were studied and the genetic association of selection traits important to pulpwood and solid wood breeding objectives were examined. These base population trials comprised of 135 families derived from native stand seed collections of the 13 races from which the Australian National E. globulus Breeding Population was founded. Significant additive genetic variation was found for all traits (stem diameter at breast height [DBH], stem straightness, acoustic wave velocity, wood basic density and pulp yield). There was no adverse race or additive level genetic correlation of DBH with any of the other traits studied. While race and additive genetic correlations were usually aligned, significant opposing genetic correlations were evident at these different genetic scales for pulp yield and wood basic density. Furthermore, key breeding objective-specific traits were either favourably (pulp yield - acoustic wave velocity) or not significantly (pulp yield - straightness) genetically correlated, arguing that genetic improvement between pulpwood and solid wood are well aligned and that breeding for one objective will have no adverse impact on the other. Wood basic density in the previous study was assessed using wood cores taken from standing trees using a mechanical corer. This is a relatively expensive assessment approach and has limited the large-scale assessment of wood density in breeding trials. A hand-held IML Power Drill series instrument (RESI) is starting to be used in the forest industry to measure wood basic density. The RESI trace also can be used to extract measures of stem diameter and bark thickness. This study quantified and confirmed the genetic association between RESI resistance values and core basic density (≥0.95). It also confirmed high genetic correlations (>0.90) of bark thickness and diameter (DBH) estimated from RESI to the analogous traditional methods. Significant family and subrace differentiation were detected for the three RESI-derived traits, with the subrace differentiation for bark thickness exhibited among the highest subrace differentiation (QST > 0.63) reported to date for E. globulus, signalling divergent selection. This study confirmed that traditional measurements of wood density, DBH and bark thickness can be replaced with RESI measurements, for the genetic studies of Eucalyptus globulus. While OP trials have been useful to demonstrate significant racial variation within E. globulus and provide the initial estimates of the levels of additive genetic variation for selection traits, they do not allow the estimation of non-additive genetic effects. This is a key issue now that the advanced generations of the Australian National E. globulus Breeding Program comprise full-sib families and there are deployment options for exploiting non-additive genetic effects through full-sib family deployment. The relative importance of non-additive genetic effects, including inter-race heterosis, was examined using a trial derived from first generation selections from the breeding program. The trial was 9 year-old and established from 515 full-sib families derived from a diallel crossing design involving intra- and inter-race hybrids of the three most widely used races in the breeding program. Growth (diameter at breast height; DBH) was assessed at ages 2, 4, 6, and 8 years, allowing the detection and monitoring of changes in additive, dominance/heterosis as well as maternal and reciprocal effects with age. Key findings include the generally insignificant maternal and reciprocal effects, significant dominance variance (22 to 34% of the additive variance) and significant inter-race heterosis which increased with age (2.2% to 6.5%). While not significant, all inter-race combinations were better than the best of their intra-race crosses (‘better-parent’ heterosis), consistent with low levels of inbreeding in intra-race crosses. Three replicates of the trial were non-destructively assessed for pulp yield using NIR spectroscopy, and resistance drilling (RESI) used to assess wood basic density as well as bark thickness. The reliability of the RESI basic density estimates were validated, and in contrast to growth, these traits were shown to be predominantly under additive genetic control. In summary, the studies of both open-pollinated (OP) and control-pollinated progeny trials have provided novel insights into the genetic architecture of growth, wood property traits and bark thickness in Eucalyptus globulus. The demonstration of significant non-additive genetic effects for growth demonstrates the advantage of identifying the best heterotic full-sib families for deployment through mass-supplementary pollination, which also gives the additional benefit of avoiding inbreeding. In addition, the study confirms the neutral or favourable association between pulpwood and solid wood traits in this species, arguing that with the breeding so far focused on pulpwood, pulpwood-selected germplasm and current plantations will not be degraded in terms of their genetic suitability to use them for solid-wood products