| Progeny testing plays a key role in the process of genetic improvement of forest trees.It can provide phenotypic information on important economic traits,and is the basis for estimating genetic parameters and selecting elite parents,families,and offspring individuals.The quality of progeny measurement is directly related to whether the genetic gain can be improved and whether the breeding work can achieve sustainable development.How to improve the prediction accuracy of the assay and maximize the efficiency of genetic improvement has always been one of the priorities of breeders.At present,with the exception of a few tree species that have entered the advanced generations,genetic determination of forest trees still remains in the offspring determination based on open pollination materials,and conventional experimental design is difficult to control the environmental heterogeneity in forests.At the same time,the competition effects between individuals are ignored,resulting in inaccurate genetic parameter estimates and other issues,thereby reducing the efficiency of genetic improvement.Therefore,in this paper,with the Japanese larch trail established in 1988,which consisted of 55 full-sibs and 94 open-pollinated families with 12×12 full-diallel matings,we conducted analysis of combining ability and reciprocal effects for growth and wood chemical properties using full-sib family materials.Using pedigree reconstruction technique to restore paternal information of open-pollinated progeny and study the effects of pedigree reconstruction,spatial and competition effects on genetic parameter estimates and breeding value rankings.Finally,a general procedure for progeny testing was systematically proposed to provide a scientific basis for the development of Japanese larch breeding strategies.Main results were showed as follows:(1)Both general combining ability(GCA)and special combining ability(SCA)had significant effects on the growth traits.SCA was dominant at 19-year-old,and its importance was decreased at 28-year-old.More genetic gain would be captured by the utilization of SCA through producing improved seeds from mating those parent pairs with high GCA and highSCA effects.Further,genetic gain could be maximized through vegetative propagation of the trees developed from those seeds.The wood properties were more strongly controlled by genetic effects that growth traits,and the individual heritabilities were greater than 0.3,with lignin being the highest,0.7.There was a significant positive and negative correlation between growth traits and lignin and cellulose content,respectively.The reciprocal effect for growth traits was not significant except for tree height at age 28,while were significant for wood chemical,indicating the mating direction could be ignored for growth traits but not for chemical triats.The majority(75%)of the parents did not show significant self-depression,and there was no significant difference between the survival of selfing and non-selfing offsprings.(2)The additive genetic variances and heritabilities of were decreased for all traits after genetic reconstruction.The dominant effect significantly affected the tree height,-cellulose and lignin content.The estimated accuracy of breeding value and realized genetic gain is higher than that before pedigree reconstruction,while the realized genetic gain declined due to the decrease of additive variance.The genetic gain decreased 26%,40%,56%,42% and 51%for DBH,volume,-cellulose,lignin(low content is superior)and holocellulose when forward selection(progeny selection)was made at 5% selection intensity.(3)The wood properties are less affected by the environment,spatial effects are not significant,and there are obvious spatial heterogeneities in growth traits.The first-order autoregressive(AR)model can eliminate most of the plot effect variance(>80%)and significantly increase the heritability estimates for H19,H28,DBH28,and V28(17.6%-41.7%).The spatial model had little effect on parental and offspring breeding value accuracy.The accuracy of parental and progeny breeding values(0.73 and 0.85)were higher than those of growth traits(0.50-0.59).(4)The competition effect was introduced into the model,which significantly increased the direct additive genetic variance()(205% and 93%)for DBH and volume,and reduced the residual variance()(8% and 6%).In the model,competition and spatial effects were introduced at the same time.The estimates of increased by 6% and 27% respectivelycompared to the competition model,and the estimates of decreased by 5% and 9%,respectively.The estimates of direct competition genetic variance()increased by 23% and4%,respectively.Competition and spatial effects have a great influence on selection.When selected with the intensity of 5%,the proportions of the joint individuals of the basic model and the competition model and the competitive spatial model are only 39.1% and 24.7%respectively,indicating that ignoring these two effects will affect the efficiency of genetic improvement. |
PDF Full Text Request