Haplotype-resolved regulation of secondary metabolism in cannabis

Cannabis sativa produces a rich array of specialized metabolites, including cannabinoids, terpenoids, and flavonoids, which are of immense therapeutic interest. However, the genetic and regulatory complexity arising from its highly heterozygous genome has obscured a complete understanding of their biosynthesis. Here, we present a fully phased, chromosome-level genome assembly of the wild diploid cannabis accession CSLZ. Comparative genomics and transcriptomics reveal that extensive haplotype-specific variation is a major driver of metabolic diversification within this accession. We find that structural variants and transposable element insertions differentially shape the allelic architecture of key biosynthetic genes, leading to divergent expression and function between haplotypes. This is exemplified by haplotype-specific loss of germacrene D synthase activity, copy number variation in oxidosqualene cyclases, and silencing of a flavonoid glycosyltransferase allele via promoter motif loss. Furthermore, allele-specific expression contributes to the spatiotemporal regulation of metabolic pathways, from cuticle formation in leaves to cannabinoid production in glandular trichomes. Our haplotype-resolved resource uncovers the pervasive role of allelic divergence in generating the specialized metabolome of wild cannabis accession CSLZ, providing a foundation for genomics-guided breeding and metabolic engineering.