Infinium iselect snp array

DNA for the 4,245 trees was isolated from needles and twig buds with the Qiagen DNeasy Plant Kit and quantified with PicoGreen fluorescent dye (Invitrogen). Trees were genotyped using an Infinium iSelect SNP array (Illumina, San Diego, CA) as described in Lenz et al. (2020a) (link). After applying several filters for retaining high-quality data, imputing only a small proportion of genotypes (0.9%), and correcting the registered pedigree using marker data (see Suppl. Methods), a total 4,066 trees from 146 families, genotyped on 4,092 SNPs, were retained for quantitative genetic analyses. The retained SNPs had an average call rate of 99.1%, an average genotyping reproducibility rate of 99.99% as assessed by replicated control genotypes, an average minor allele frequency MAF = 0.211, and an average fixation index F_e = 0.022.
Genetic diversity within each BG was estimated with the status effective number (Ns):
where θ is the group coancestry (Lindgren et al., 1996 ), as estimated from the corrected full-sib pedigree. Descriptive statistics of BGs and phenotypes are presented in Tables 1, 2, respectively. Boxplots of phenotypes grouped by sites and BGs are shown in Figure S2.

The 726 trees were genotyped using an Infinium iSelect SNP array (Illumina), which was assembled from a catalog of high‐confidence gene SNPs obtained from exome capture and sequencing (Azaiez et al., 2018). The genotyping reproducibility rate was high (99.94%), as estimated from two positive controls replicated on each genotyping plate. From 5,660 successfully manufactured SNPs representing as many distinct gene loci well distributed over the 12 spruce linkage groups (Pavy et al., 2017), we retained a total of 3,914 SNPs with call rate ≥90% (average call rate of 99.6%), minor allele frequency (MAF) ≥0.005, and a fixation index |F_IS| < 0.50. SNPs were well distributed across MAF classes with 86% of the SNPs with MAF ≥0.05 (Figure S4). Missing genotypes (only 0.4% of genotypes) were imputed using a k‐nearest neighbor method based on linkage disequilibrium (LD‐kNNi) with the software LinkImpute (Money et al., 2015). The software estimated an accuracy of 0.83 for imputed genotypes by randomly masking genotypes.

Genomic profiles were obtained for each tree from genotyping with an Infinium iSelect SNP array (Illumina, California) described in Lenz, Nadeau, Azaiez, et al. (2020 (link)). After quality filters, data from 4091 high‐quality Mendelian single‐nucleotide polymorphisms (SNPs) with little or no missing data and representing as many distinct gene loci well dispersed over the white spruce genome (Lenz, Nadeau, Azaiez, et al., 2020 (link)) were used for GS analyses. Pedigree informed by paternity recovery was also obtained from Lenz, Nadeau, Azaiez, et al. (2020 (link)). Paternal assignment and pedigree verification were conducted so that an informed and corrected pedigree information was also available for conventional ABLUP analyses. In doing so, a total of 347 genetically distinct full‐sib families implicating 38 mothers and 19 fathers could be recovered, in agreement with the 38 families obtained by polycross mating and the 19 fathers used in the pollen polymix to sire the female trees.