Blueberry Pan-Genome Reveals Hidden Diversity in Sugar-Tagging Enzyme Genes

The deep blue hue of a ripe highbush blueberry is not simply a matter of pigment production; it is the result of an elaborate chemical finishing process in which sugars are attached to anthocyanins, flavonols and other specialized metabolites by a large family of enzymes known as UDP-glycosyltransferases, or UGTs. These enzymes dictate not only the color and stability of the fruit’s signature compounds but also its flavor profile, nutritional value and resistance to environmental stress. Yet despite their central role in shaping one of the world’s most celebrated superfoods, the true genetic diversity of the UGT superfamily across cultivated blueberry germplasm has remained poorly resolved. A new study published in BMC Genomics now offers the most comprehensive picture to date, using a pan-genome approach to reveal that the blueberry UGT repertoire is far larger, more variable and more evolutionarily dynamic than any single reference genome could suggest.

Researchers Bin Li, Zhuqing Zheng and Ruiyi Fan of Jingchu University of Technology in Jingmen, China, constructed a UGT pan-genome spanning 23 cultivars of highbush blueberry, Vaccinium corymbosum. Their analysis identified a striking total of 11,696 VcUGT genes distributed across 662 orthologous gene groups, or OGGs. This substantially extends the previously cataloged inventory of 361 VcUGT members, which had been derived from the single ‘Draper’ reference genome. The finding underscores a lesson that is rapidly becoming a central theme of modern plant genomics: a single reference assembly, however high quality, captures only a fraction of the functional gene diversity present within a crop species, particularly one with a complicated polyploid genome like the autotetraploid highbush blueberry.

The structure of the pan-genome itself proved revealing. The researchers found it to be highly open, meaning that each additional cultivar sequenced continues to contribute new gene variants rather than quickly saturating the inventory. At the heart of the pan-genome sits a compact core of just 106 orthogroups, which nevertheless accounts for 57.5 percent of all gene copies, reflecting how frequently these conserved UGT lineages are duplicated within individual genomes. Surrounding that core is a remarkably plastic dispensable fraction of 556 orthogroups, including 374 so-called cloud orthogroups that are present in only a minority of cultivars. Notably, 328 of these cloud groups were cultivar-specific, meaning they exist in just one variety among the 23 examined, a level of divergence that hints at extensive gene birth, loss and reshaping during the recent history of blueberry breeding.

One of the most striking evolutionary patterns emerged when the blueberry UGT families were compared with those of Arabidopsis thaliana, the standard reference plant for genomic comparisons. Subfamily G, corresponding to the UGT85 lineage, showed an 18-fold expansion relative to Arabidopsis. Because highbush blueberry is an autotetraploid, carrying four copies of each ancestral chromosome set, the expansion translates to roughly 4.5-fold on a per-haploid-genome basis, still a dramatic proliferation. At the same time, clades involved in sterol metabolism and stress responses underwent reciprocal contraction, suggesting that during the evolutionary history of the Vaccinium lineage, the enzyme family was sculpted in a direction that favors the glycosylation of flavonoid-type compounds, the very chemistry that gives blueberries their characteristic pigments and antioxidant-rich composition.

To probe the selective forces acting on these genes, the team performed pairwise analyses of nonsynonymous to synonymous substitution rates, the classic Ka/Ks test, and examined the PSPG box, a highly conserved motif within the plant secondary product glycosyltransferase domain that plays a crucial role in recognizing the sugar donor substrate. The results indicated pervasive purifying selection acting on the catalytic core of these enzymes, consistent with the idea that the fundamental chemistry of sugar transfer is too essential to tolerate much tinkering. Natural selection appears to preserve the enzymatic machinery while allowing variation to accumulate in the regulatory and copy-number dimensions of the family.

That interpretation gained further support from the analysis of structural variants. Structural variants, encompassing insertions, deletions and inversions of substantial DNA segments, are increasingly recognized as major drivers of crop trait diversity, and the blueberry UGT pan-genome proved no exception. Pangenes affected by structural variants showed elevated paralog divergence specifically within the dispensable fraction of the genome, a pattern the authors interpret as consistent with relaxed rather than diversifying selection. In other words, freed from the strong functional constraints that govern the core gene set, dispensable UGT copies accumulated mutations and diverged from one another, generating raw material on which future selection, whether natural or breeder-directed, could potentially act.

In a display of methodological caution that stands out in an era of genome-scale enthusiasm, the researchers identified two extreme-length gene models belonging to the SNAP category and classified them as suspect or fused candidates. Rather than attempting to draw mechanistic conclusions from these anomalous models, they excluded them from functional interpretation pending transcript-level validation. This restraint matters because spurious gene predictions, particularly in complex polyploid assemblies where duplicated haplotypes can be misassembled into fused models, can mislead downstream analyses and breeders alike. By flagging and setting aside such cases, the study models a rigorous standard for pan-genome research in polyploid crops.

The comparative dimension of the work also delivered practical insights for breeding programs. Northern highbush and southern highbush blueberries represent two major cultivated groups, distinguished historically by their adaptation to cold-winter and mild-winter climates respectively, with southern highbush varieties developed by incorporating native Vaccinium species adapted to warmer conditions. When the researchers compared the two groups, they identified 15 orthogroups with significant group-biased presence between northern and southern highbush cultivars, and all 15 were confined to the dispensable fraction of the pan-genome. This pattern suggests that the genetic distinctions underlying cultivar-group differentiation include dispensable, structurally variable UGT loci, raising the intriguing possibility that variation in these glycosylation genes contributes to group-specific fruit quality characteristics or stress adaptation traits.

For breeders, the implications are direct and tangible. Anthocyanin and flavonol glycosylation influences the intensity and stability of fruit color, the accumulation of health-associated polyphenols, and potentially the shelf life and processing quality of berries, all of which are commercial priorities. The pan-genome catalog produced by this study provides a roster of SV-associated and dispensable UGT loci that now constitute candidate targets for functional validation and marker-assisted breeding. Rather than relying on a single reference genotype, breeders seeking to modulate polyphenol profiles or fruit pigmentation can consult a diversity-aware inventory that reflects the actual breadth of cultivated germplasm. As genomic selection and precision breeding techniques advance in horticultural crops, having an accurate map of which enzyme-coding families vary across varieties and which are anchored in the conserved core becomes a powerful navigational tool.

More broadly, the study adds blueberry to the growing list of crops, ranging from wheat and maize to tomato and potato, for which pan-genome analyses have exposed the limitations of reference-centric genomics. In polyploid species especially, where multiple haplotypes coexist within every individual, structural variation and presence-absence variation can hide entire gene families from view. The 11,696 VcUGT genes identified across just 23 cultivars, compared with the 361 cataloged from the ‘Draper’ reference, represent a more than thirtyfold difference in copy count, a vivid illustration of how much biology remains concealed within unsequenced diversity. As sequencing costs continue to fall and pan-genome frameworks mature, the blueberry UGT catalog will likely serve as a template for exploring other metabolically important gene families in Vaccinium and related ericaceous crops, connecting genome-scale discovery to the sensory and nutritional qualities that consumers experience with every handful of berries.

Subject of Research: Pan-genomic analysis of UDP-glycosyltransferase gene diversity and structural variation across highbush blueberry cultivars

Article Title: Structural variation and pan-genome dynamics shape the UDP-glycosyltransferase repertoire in highbush blueberry

Article References: Li, B., Zheng, Z., & Fan, R. (2026). Structural variation and pan-genome dynamics shape the UDP-glycosyltransferase repertoire in highbush blueberry. BMC Genomics. https://doi.org/10.1186/s12864-026-13376-5

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13376-5

Keywords: Vaccinium corymbosum, UDP-glycosyltransferase, pan-genome, structural variation, highbush blueberry, anthocyanins, flavonoid glycosylation, orthogroups, polyploid genome, marker-assisted breeding, gene family expansion, presence-absence variation