Pangenome of U.S. ex-PVP and Wild Sorghum Reveals Structural Variants and Selective Sweeps Shaping Adaptation and Trait Improvement

Abstract

Sorghum is a cereal crop grown for food, feed, and biofuel, yet the genomic diversity of elite commercial lines remains underexplored. Here, we assembled a sorghum pangenome from long-read sequences of 46 U.S. ex-Plant Variety Protection Act (ex-PVP) cultivars and a haplotype-resolved assembly of one wild accession. The pangenome revealed largely conserved genomic architecture across elite lines but identified presence-absence variations (PAVs) in 28% of gene families, including genes involved in stress responses and starch metabolism, such as STARCH BRANCHING ENZYME I (SBE1). Selective sweep analyses detected strong signals at key flowering-time loci (Ma1/SbPRR37, Ma2, Ma6) and circadian/light regulators (PHOT1, DET1, XAP5), with PHYA showing copy number variations (CNVs). Analysis of 24 wild accessions, beyond the core pangenome panel, detected CNVs in MULTIDRUG AND TOXIC COMPOUND EXTRUSION (MATE) transporters, potentially reflecting selection or breeding for enhanced herbicide detoxification. These results highlight breeding-shaped regions and position the pangenome as a resource for sorghum improvement.