Development of a highly efficient 50K single nucleotide polymorphism genotyping array for the large and complex genome of Norway spruce (Picea abies L. Karst) by whole genome resequencing and its transferability to other spruce species

Abstract

Norway spruce (Picea abies L. Karst) is one of the most important forest tree species with significant economic and ecological impact in Europe. For decades, genomic and genetic studies on Norway spruce have been challenging due to the large and repetitive genome (19.6 Gb with more than 70% being repetitive). To accelerate genomic studies, including population genetics, genome-wide association studies (GWAS) and genomic selection (GS), in Norway spruce and related species, we here report on the design and performance of a 50K single nucleotide polymorphism (SNP) genotyping array for Norway spruce. The array is developed based on whole genome resequencing (WGS), making it the first WGS-based SNP array in any conifer species so far. After identifying SNPs using genome resequencing data from 29 trees collected in northern Europe, we adopted a two-step approach to design the array. First, we built a 450K screening array and used this to genotype a population of 480 trees sampled from both natural and breeding populations across the Norway spruce distribution range. These samples were then used to select high-confidence probes that were put on the final 50K array. The SNPs selected are distributed over 45,552 scaffolds from the P. abies version 1.0 genome assembly and target 19,954 unique gene models with an even coverage of the 12 linkage groups in Norway spruce. We show that the array has a 99.5% probe specificity, >98% Mendelian allelic inheritance concordance, an average sample call rate of 96.30% and an SNP call rate of 98.90% in family trios and haploid tissues. We also observed that 23,797 probes (50%) could be identified with high confidence in three other spruce species (white spruce [Picea glauca], black spruce [P. mariana] and Sitka spruce [P. sitchensis]). The high-quality genotyping array will be a valuable resource for genetic and genomic studies in Norway spruce as well as in other conifer species of the same genus.

Keywords: Norway spruce; SNP array; genetic diversity; genome resequencing; genomic selection.

FIGURE 1

Schematic illustration of the variant filtering pipeline for extracting candidate probe sequences for the Axiom in silico testing at ThermoFisher. Each of the filtering steps described in the text is presented in a grey boxes with the number of surviving SNPs labelled beside

FIGURE 2

Visualization of the additive relatedness matrix estimated across all 468 samples. The relatedness matrix was calculated with the A.mat function in the R package “rrBLUP” using all PolyHigh resolution SNPs (176,800). Inset: zoom of the nine replicated samples

FIGURE 3

Schematic illustration of the probe selection pipeline from the 450K screening array to the final 50K array

FIGURE 4

Scatter plot of the minor allele frequency and heterozygosity for the final SNP selection (50K, right red) in comparison to all screened SNPs (450K, grey) and all PolyHigh resolution SNPs (177K, dark red)

FIGURE 5

Population structure estimated using a principal component analysis on the relatedness matrix calculated based on all 177K PolyHigh resolution SNPs (top row) and from the final 50K SNP selection (bottom row). Left‐hand panels are coloured based on which provenance trial the samples origginate from while the right‐hand panels are coloured based on documented sample origin. Replicated samples have been removed from the analysis. NFE—Fennoscandia contains samples from Finland and northern Sweden; C‐sc—Southern Scandinavia from Central/Southern Sweden and Central/Southern Norway; Rus_Bal—Russian‐Baltic from Russia, Belarus, Estonia, Latvia and Lithuania; NPL—Northern Poland; ROM—Carpathian from Romania and Bulgaria; CEU—Central Europe from Slovakia, Czech Republic, Southern Poland, Hungary and Austria; ALP—Alpine from Denmark, Germany, Switzerland, France and Italy; U—unknown

FIGURE 6

Summary of the array evaluation metrics. (a) Histogram of the sample call rate for Norway spruce. The dashed red line indicates the averaged call rate. (b) Histogram of the probe call rate for Norway spruce. The dashed red line indicates the averaged call rate. (c) Histogram of the proportion of homozygous calls for 45,541 probes estimated using 24 haploid tissues. The dashed red line indicates the averaged proportion of homozygous calls. (d) Histogram of the Mendelian inheritance (MI) error rate for 36,256 probes estimated using 48 family trios. (e) Principal component analysis for all four spruce species. (f) Principal component analysis for the three non‐Norway spruce species