W309 Twenty Years of Eucalyptus Molecular Breeding: From RAPD and QTLs to Genotyping-By-Sequencing and Genomic Selection of Complex Traits

Date: Sunday, January 15, 2012
Time: 5:50 PM
Room: Sunrise
Dario Grattapaglia , Plant Genetics Laboratory, EMBRAPA Genetic Resources and Biotechnology, Brasilia, DF, Brazil
At the first PAG Conference in 1992 we presented the first Eucalyptus linkage maps and the prospects of using them for Marker Assisted Selection (MAS), a possibility that became the cornerstone of several forest tree genome projects in subsequent years. Twenty years have gone by however, and MAS has not made it yet in the "real tree breeding world", despite the large volume of QTL and association genetics (AG) data published. Reasons include limitations of early genomic technologies, the rapid decay of LD and an overoptimistic outlook of the architecture of complex traits in forest trees. QTLs effects are typically overestimated while candidate-gene AG captures small fractions of trait heritability. The advent of high-throughput genotyping technologies coupled to genome-wide prediction of phenotypes by Genomic Selection (GS) have provided a new paradigm to integrate genomics into breeding. In four independent Eucalyptus breeding populations we have now shown that GS accuracies (0.55-0.88) matched or surpassed conventional phenotypic selection and substantial proportions (74%-97%) of trait heritability were captured for traits such as volume growth, wood specific gravity, pulp yield and rust resistance. GS accuracies across populations were however poor, likely due to variable allelic effects, inconsistent patterns of LD and GxE. Genotyping-by-sequencing technologies will likely be the workhorse of GS, while strategic and economic aspects of the operational adoption of GS will receive increased attention by breeders. GS opens a new perspective to a genome-wide understanding of quantitative trait variation in forest trees and shall realize the standing promises of genomics into applied tree breeding.