Pangenome and pan-transcriptome analyses reveal evolutionary diversification of the GPAT gene family in soybean

Glycerol-3-phosphate acyltransferases (GPATs) participate in plant lipid metabolism, yet their intraspecific evolutionary and transcriptional diversity in soybean remains incompletely characterized. Here, we integrated a survey of 1,011 plant species with pangenome and pan-transcriptome analyses of 29 soybean accessions. We identified 15,799 GPAT genes across plants and 1,329 across soybean accessions; the latter were classified into three subfamilies, with 42-47 genes per accession. Among soybean GPAT genes, 1,317 were assigned to 47 orthogroups, comprising 19 core and 28 dispensable groups, of which 22 were softcore. Whole-genome/segmental duplication accounted for 86.38% of soybean GPAT genes, and 94.04% of homologous gene-pair comparisons exhibited Ka/Ks < 1, indicating predominant purifying selection. Whole-genome/segmental duplicates were more prevalent among core genes, which also showed lower median Ka/Ks ratios than dispensable genes. Subfamily III contained the highest proportion of dispensable genes, whereas subfamily II exhibited the highest mean Ka/Ks ratio, revealing distinct patterns of gene-content and coding-sequence divergence. Expression profiling across eight tissues revealed transcriptional divergence within subfamilies and greater expression stability of core pangenes across accessions. Co-expression networks shared 27 nodes and 25 edges among cultivars, landraces, and wild soybeans, with both core and dispensable members occupying highly connected positions. Together, these results highlight the coexistence of evolutionary conservation, gene-content variation, and transcriptional divergence within the soybean GPAT family. These findings provide candidate GPAT genes and testable hypotheses for future functional studies related to soybean lipid traits and stress tolerance.