Genomic selection in forest tree breeding - basic principles, problems and future prospects

All tree breeders cope with the same challenge of the very long time interval of a single breeding cycle. What is more, trees are long-lived, with desirable breeding traits expressing late during their life cycle. Increasing problems with climate change, globalization or economic growth have forced us to accelerate tree breeding and improve selection precision, both of which can be achieved by genomic selection (GS). The idea of GS was introduced nearly 20 years ago as an extension of marker-assisted selection (MAS) in order to advance breeding technologies using genetic markers. Unlike MAS, which exploits only a set of marker-trait associations, GS relies on a high number of genetic markers that are spread throughout the entire length of the genome. All markers effects are assessed simultaneously in order to build a precise model that allows prediction of genetic estimated breeding value of a particular individual using genetic data only. GS has already revolutionized dairy cattle breeding resulting in remarkable improvements across multiple traits and is becoming more and more common in crop production. We now know that genetic architecture of quantitative traits is complex, but recent advances in genomics have made it possible to deal with this problem in an unprecedented way. There are certain concerns regarding GS in forest tree species that include genotype-environment (GxE) interaction and the usefulness of the predictive model built up by GS in the next generation of trees. Nevertheless, experimental results obtained so far have shown that the genetic gain per unit time as well as selection precision can be substantially increased. Here we present the basic principles of GS for forest tree species, giving examples of studies carried out so far and discussing problems and future possibilities that GS may soon open up for forest tree breeders.