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

Abstract Background UDP-glycosyltransferases ( UGTs ) shape blueberry ( Vaccinium corymbosum ) fruit quality by glycosylating anthocyanins, flavonols and other specialized metabolites, yet the genetic diversity of this superfamily across cultivated germplasm remains unresolved. Results Here, we constructed a UGT pan-genome across 23 highbush blueberry cultivars, identifying 11,696 VcUGT genes in 662 orthogroups (OGGs) and substantially extending the 361 VcUGT members previously cataloged from the single ‘Draper’ reference. The pan-genome is highly open, comprising a compact 106-OGG core (57.5% of all gene copies) and a plastic 556-OGG dispensable fraction including 374 cloud OGGs, of which 328 were cultivar-specific.Subfamily G ( UGT85 ) shows an 18-fold expansion relative to Arabidopsis (~ 4.5-fold per haploid equivalent given the autotetraploid genome), coupled with reciprocal contraction of sterol- and stress-related clades. Pairwise Ka / Ks and PSPG-box analyses indicate pervasive purifying selection on the catalytic core, while structural-variant-affected pangenes show elevated paralog divergence in the dispensable fraction, consistent with relaxed rather than diversifying selection. Two extreme-length SNAP gene models were classified as suspect/fused candidates and excluded from mechanistic interpretation pending transcript-level validation. Pan-genome comparison further identifies 15 group-biased orthogroups between northern and southern highbush cultivars, all confined to the dispensable fraction. Conclusions Collectively, the pan-genome catalogue identifies SV-associated and dispensable UGT loci for future functional validation and marker-assisted breeding, while avoiding mechanistic interpretation of unvalidated gene models.