Evolutionary Applications (Dec 2020)

Genomic selection for resistance to spruce budworm in white spruce and relationships with growth and wood quality traits

  • Jean Beaulieu,
  • Simon Nadeau,
  • Chen Ding,
  • Jose M. Celedon,
  • Aïda Azaiez,
  • Carol Ritland,
  • Jean‐Philippe Laverdière,
  • Marie Deslauriers,
  • Greg Adams,
  • Michele Fullarton,
  • Joerg Bohlmann,
  • Patrick Lenz,
  • Jean Bousquet

DOI
https://doi.org/10.1111/eva.13076
Journal volume & issue
Vol. 13, no. 10
pp. 2704 – 2722

Abstract

Read online

Abstract With climate change, the pressure on tree breeding to provide varieties with improved resilience to biotic and abiotic stress is increasing. As such, pest resistance is of high priority but has been neglected in most tree breeding programs, given the complexity of phenotyping for these traits and delays to assess mature trees. In addition, the existing genetic variation of resistance and its relationship with productivity should be better understood for their consideration in multitrait breeding. In this study, we evaluated the prospects for genetic improvement of the levels of acetophenone aglycones (AAs) in white spruce needles, which have been shown to be tightly linked to resistance to spruce budworm. Furthermore, we estimated the accuracy of genomic selection (GS) for these traits, allowing selection at a very early stage to accelerate breeding. A total of 1,516 progeny trees established on five sites and belonging to 136 full‐sib families from a mature breeding population in New Brunswick were measured for height growth and genotyped for 4,148 high‐quality SNPs belonging to as many genes along the white spruce genome. In addition, 598 trees were assessed for levels of AAs piceol and pungenol in needles, and 578 for wood stiffness. GS models were developed with the phenotyped trees and then applied to predict the trait values of unphenotyped trees. AAs were under moderate‐to‐high genetic control (h2: 0.43–0.57) with null or marginally negative genetic correlations with other traits. The prediction accuracy of GS models (GBLUP) for AAs was high (PAAC: 0.63–0.67) and comparable or slightly higher than pedigree‐based (ABLUP) or BayesCπ models. We show that AA traits can be improved and that GS speeds up the selection of improved trees for insect resistance and for growth and wood quality traits. Various selection strategies were tested to optimize multitrait gains.

Keywords