Unveiling the evolutionary history of lingonberry (Vaccinium vitis-idaea L.) through genome sequencing and assembly of European and North American subspecies

Abstract

Lingonberry (Vaccinium vitis-idaea L.) produces tiny red berries that are tart and nutty in flavor. It grows widely in the circumpolar region, including Scandinavia, northern parts of Eurasia, Alaska, and Canada. Although cultivation is currently limited, the plant has a long history of cultural use among indigenous communities. Given its potential as a food source, genomic resources for lingonberry are significantly lacking. To advance genomic knowledge, the genomes for 2 subspecies of lingonberry (V. vitis-idaea ssp. minus and ssp. vitis-idaea var. 'Red Candy') were sequenced and de novo assembled into contig-level assemblies. The assemblies were scaffolded using the bilberry genome (Vaccinium myrtillus) to generate a chromosome-anchored reference genome consisting of 12 chromosomes each with a total length of 548.07 Mb [contig N50 = 1.17 Mb, BUSCO (C%) = 96.5%] for ssp. vitis-idaea and 518.70 Mb [contig N50 = 1.40 Mb, BUSCO (C%) = 96.9%] for ssp. minus. RNA-seq-based gene annotation identified 27,243 and 25,718 genes on the respective assembly, and transposable element detection methods found that 45.82 and 44.58% of the genome were repeats. Phylogenetic analysis confirmed that lingonberry was most closely related to bilberry and was more closely related to blueberries than cranberries. Estimates of past effective population size suggested a continuous decline over the past 1-3 MYA, possibly due to the impacts of repeated glacial cycles during the Pleistocene leading to frequent population fragmentation. The genomic resource created in this study can be used to identify industry-relevant genes (e.g. anthocyanin production), infer phylogeny, and call sequence-level variants (e.g. SNPs) in future research.

Keywords: V. vitis-idaea ssp. minus; Vaccinium vitis-idaea ssp. vitis-idaea; genome assembly; lingonberry; mountain cranberry; partridgeberry.

Fig. 1.

a) Worldwide distribution of V. vitis-idaea L. (GBIF 2023). Dots represent occurrence records registered as follows: V. vitis-idaea ssp. minus (blue), V. vitis-idaea ssp. vitis-idaea (red), and V. vitis-idaea L. ssp. unidentified (yellow). b) V. vitis-idaea ssp. vitis-idaea flowers and fruits. c) V. vitis-idaea ssp. minus (left) and ssp. vitis-idaea var. ‘Red Candy’ (right) grown in the greenhouse.

Fig. 2.

a) Gene and TE distributions in lingonberry genome (V. vitis-idaea var. ‘Red Candy’). b) Gene and c) TE densities by distance from centromeres. Centromere positions were approximately mapped from bilberry genome as a range, and distance was calculated to its middle value (Wu et al. 2021). Red shades indicate the gene density, and purple shades indicate the TE density. Genes were filtered to represent only the longest gene in case of isoforms and splicing variants present. All densities are presented as the number of feature counts per 1 Mb, except the terminal windows <1 Mb.

Fig. 3.

Heatmap of gene abundance related to phenolic compound biosynthesis. Columns represent sample type, and rows represent gene copies on lingonberry genome. Sample types are (from left to right) V. vitis-idaea var. ‘Red Candy’ rhizome, leaf, flower, berry, and var. ‘Sunna’ berries at different ripening stages; green berry, white berry, and red berry (Tian et al. 2020). Abundance was measured by FPKM. Note that the red color gradient was normalized within each heatmap, so comparison cannot be made across heatmaps. The enzyme pathway is based on Colle et al. (2019).

Fig. 4.

a) Past effective population size (N_e) of lingonberry with MSMC2. The N_e of V. vitis-idaea ssp. minus (blue; LW1) and V. vitis-idaea ssp. vitis-idaea var. ‘Red Candy’ (red; LC1) was plotted against years before present. Both x- and y-axes were log scaled. Plots were generated with the generation time of 5 years and mutation rate of 3 × 10⁹ mutations/generation. Note the timings are presented as the range estimate from generation time of 5–10 years. b) Phylogeny of Vaccinium based on 2,226 conserved BUSCO genes. Thick lines indicate nodes supported by >60 STAG support values in OrthoFinder (Emms and Kelly 2018, 2019). The numbers on the selected node represent divergence time in million years (MY), calibrated at the divergence time with Rhododendron (45.5–76.9 MY), and the number in bracket shows the gene concordance factor (0–100) obtained from 2,226 BUSCO genes.