Abstract:

Genomic selection was supposed to be the key technology and hotspot of those studies on animal and plant genetic breeding. Some progress has been made toward the genetic improvement of some animal and plant species. The breeding cycle of tree species is long due to their long generation interval. Early selection is an effective way to shorten the breeding cycle of trees and speed up the breeding process. The studies of early selection of trees can be roughly divided into three stages, including early selection based on the correlation between the early and late trait values, molecular marker-assisted selection and genomic selection. The growth and wood quality traits, the main target traits for tree genetic improvement, are often complex traits controlled by quantitative traits. During the constant process of tree growth and development, these traits are suffered from the on-going influence of the environment. Meanwhile, the heritability of these traits may be gradually changing in the process of tree growth and development. However, the application of genomic selection in tree genetic improvement was limited by the unique characteristic and weak research basis of perennial tree species, such as long generation interval, large body size, long juvenile phase, the lack of multi-omics data (genomic and phenotypic data), and the imperfection of technology platform to facilitate tree researches. To speed the application process of genomic selection in forest genetic improvement, this paper attempts to provide a short overview of the principles and method of whole genomic selection, to summarize recent progress of genome-wide selection in forestry tree genetics and breeding, and to discuss the constraint factors which impact the application of genomic selection in forestry tree genetic improvement. The paper provides a brief introduction of the linear model, statistical estimation methods (SNP-BLUP, GBLUP & Bayesian) and analytical tools (rrBLUP, synbreed, BGLR, GVCBLUP, GAPIT, sommer, BLUPGA and so on) used for genomic selection. It summarizes the advantage of genomic selection applied in forestry tree breeding and the studies of using genomic selection in both broadleaf Populus, Eucalyptus, Elaeis & Hevea) and coniferous (Pinus & Picea) tree species, and dissects these reports of genomic selection in Elaeis guineensis as typical cases. These studies of genomic selection in forestry tree species show that genomic selection contributes to improving the efficiency of forest selection and to accelerating tree breeding. It discusses in detail the possible influence of the tree reference genome, the genome-wide association study, the breeding population, the linkage disequilibrium and the perennial attribute on the genomic selection of trees. Genomic selection has potential application prospects in tree genetics and breeding research. Nevertheless, its feasibility still needs to be further evaluated using vast amounts of simulated and real data. The current research on tree genomic selection is facing multiple challenges: 1) a relatively low quality of genome assembly for tree species, 2) genomic selection for simultaneously screening multiple target traits, 3) the designing of the rational experimental scheme according to the characteristics of the perennial tree species, and the development of the statistical model and analytical tool capable of analyzing the longitudinal character data.

Key words: genomic selection, breeding efficiency, tree genetic improvement, genome-wide association study