2.1. Genetic material and phenotyping

The data analyzed in this study are a subset of a larger breeding population derived from a partial diallel mating design (see Mottet et al., 2015 for more details). We focused our phenotyping and genotyping efforts on the trees planted on two sites affected by white pine weevil in Quebec, that is, Saint‐Modeste (47.85°N; 69.38°W; elevation: 140 m; abbreviated STM) and Grandes‐Piles (46.68°N; 72.68°W; elevation: 150 m; abbreviated GPI), respectively, located in the balsam fir–yellow birch and the sugar maple–yellow birch bioclimatic domains (Figure ​(Figure3).3). Both tests were set up in year 2000. The Grandes‐Piles plantation was heavily affected by weevils with ~70% of the trees attacked at least once by age 16, while the Saint‐Modeste plantation was moderately affected (~47% of the trees attacked by age 16). To develop genomic selection models for weevil resistance, we selected 40 full‐sib families (35 parents), 20 of which were rated resistant and 20 rated susceptible based on weevil attack surveys at ages 10 and 15 on these two sites. A total of 726 trees (14 to 20 per family, mean = 17.85) were sampled. Each parent was crossed on average 2.3 times (Figure S1). The two trials were established according to a randomized complete block design, each with five blocks and three‐tree row plots for each family (tree spacing: 2.5 m × 2 m). Because of tree mortality, not all families were represented in every block (i.e., incomplete block design).

Location of the test sites Saint‐Modeste (STM) and Grandes‐Piles (GPI) in the province of Québec, Canada

Tree height (Height₁₅) and diameter at breast height (DBH₁₅) were measured at age 15. The height‐to‐diameter ratio (Height₁₅/DBH₁₅) was calculated as a proxy of stem taper. Three wood quality traits related to mechanical properties were assessed: average wood density and cellulose microfibril angle were determined from wood increment cores using X‐ray densitometry and diffractometry at age 15 (Density₁₅ and MFA₁₅) as described in Lenz et al. (2017); acoustic velocity was measured at age 16 (Velocity₁₆) with the ST300 Hitman tool. Acoustic velocity is a proxy for wood stiffness or modulus of elasticity measured at standing trees (Chen et al., 2015; Desponts, Perron, & DeBlois, 2017; Lenz et al., 2013).

For each of two surveys at ages 10 and 15, the presence/absence of weevil damage in the current year (coded 0/1) and in previous years (coded 0/1) were recorded. Current year damage was detected by inspecting the terminal shoot for emergence holes or death of the leader shoot. Weevil damage in previous years was visible by forks, curves, bayonets, or multiple stems. We calculated the cumulative number of attacks (CWA) as:

where WA_current10 and WA_current15 are the presence/absence of attack at age 10 and 15, respectively; WA_previous10 is the presence/absence of attack prior to age 10; and WA_11–14 is the presence/absence of attack between ages 11 and 14. The variable WA_11–14, which was calculated to avoid double‐counting previous attacks, was equal to 1 (presence) if there was an attack prior to age 15, but no attack at age 10 or prior to age 10. The resulting CWA variable is ordinal (ordered categories of 0, 1, 2, and 3 weevil attacks). Table ​Table11 and Figure S2 provide the summary statistics and violin plots, respectively, for the traits assessed in this study. Figure S3 shows the spatial distribution of CWA values in the test sites.

Phenotypic means, standard deviations (SD), and coefficients of variation (CV) for each site and across sites for the 714 trees retained for analyses