Revisiting the genome assembly of Lupinus species reveals differential diploidization after a shared whole-genome duplication

Abstract

Accurate genome assemblies are essential for comparative genomics, yet Hi-C-guided scaffolding can introduce structural errors that misrepresent chromosome architecture and bias evolutionary inferences. Here, we identified pervasive scaffolding errors - including artificial fusions, internal inversions, and incomplete contig mounting - in two previously published Lupinus genomes (L. cosentinii and L. digitatus) using a segmentation method based on LTR retrotransposon density. We re-assembled both genomes, producing chromosome-level references of 472.7 Mb (16 chromosomes) and 427.2 Mb (21 chromosomes), with BUSCO completeness >98.5%. Synteny validation and re-application of LTR profiling confirmed that all prior errors were resolved. Using these corrected genomes together with four additional Lupinus species and two outgroup legumes, we investigated post-polyploid evolution. Synonymous substitution rate (Ks) analysis revealed a genus-specific whole-genome duplication (WGD) event (Ks = 0.17) shared by all six Lupinus species. The proportion of WGD-derived genes varied markedly, from 60% in L. digitatus to only 36% in L. mutabilis, indicating differential diploidization. While all species retained a core set of WGD duplicates enriched in cytoskeleton organization, ion transport, and defense responses, each exhibited lineage-specific functional trajectories: cell wall modification in L. cosentinii and L. digitatus, nitrogen metabolism in L. albus and L. angustifolius, flower development in L. luteus, and stress/lipid metabolism in L. mutabilis. Our corrected assemblies provide optimal references for Lupinus comparative genomics, and our findings demonstrate that a shared WGD event can lead to both conserved and highly divergent post-polyploid fates, likely underpinning adaptive diversification within the genus.