Abstract
<title>Abstract</title> <p> Background Anthocyanins are the principal health-promoting pigments of blueberry ( <italic>Vaccinium corymbosum</italic> L.) and key determinants of fruit colour and nutritional value; their vacuolar accumulation requires glutathione S-transferases (GSTs) acting as cytoplasmic anthocyanin carriers, yet the genomic diversity of this family across blueberry germplasm remains unknown. Results We analysed the GST family across 23 highbush blueberry cultivars, identifying 4,304 GST genes resolved into seven canonical two-domain cytosolic GST classes and 457 orthogroups. The family showed an “open” pan-genome dominated by a large variable shell (42.0% of orthogroups; 58.6% of gene copies) and a compact core (2.6%). Tandem duplication drove the expansion of the Tau and Zeta classes, and GST loci were significantly enriched for long-read structural variants (SVs) relative to the genome average (1.39-fold; P = 8.5 × 10⁻⁶). The family was under predominant purifying selection (Ka/Ks < 1 for 93.5% of the filtered within-orthogroup gene pairs), with the anthocyanin-related Phi class the most constrained. Presence–absence and copy-number variation distinguished northern and southern highbush cultivars and was concentrated in the shell fraction. A southern highbush fruit developmental transcriptome (cv. F62) was used to prioritize GST orthogroups co-expressed with anthocyanin-biosynthetic genes; of 12 expression-supported candidates, six—including all three Phi-class members—were significant (FDR < 0.05); because Phi is the only class with functionally validated anthocyanin escorts, these three are the most credible candidates. Conclusions This study provides a pan-genome resource and identifies structurally variable, anthocyanin-co-expressed Phi-class GSTs as candidate loci for future validation of blueberry fruit pigmentation variation, with potential relevance to breeding applications. </p>