Fabaceae pan-centromere analysis traces the evolutionary origin of centromeric satellites with widespread solo-LTR transition

Abstract Centromeres are essential for chromosome segregation, but their diversity and evolution across Fabaceae remain poorly understood. Here, we present a family-wide pan-centromere dissection across 20 legume genomes using one Rosaceae genome as outgroup, incorporating our complete, gap-free assemblies of two Glycyrrhiza genomes. We dissect 251 (215 gap-free) centromeres from 18 Fabaceae members and establish a 'solo-LTR-driven' model for centromere evolution. By developing LTRtrace, a pipeline to annotate centromeric solo-LTRs, we uncover a 'molecular fossil' designated as 'satLTR' where single solo-LTRs encapsulate multiple satellite repeats. By capturing the intermediate states of LTR-to-satellite conversion, such fossil demonstrates how solo-LTRs expand into massive centromeric satellite arrays via local tandem amplification and homogenization. Finally, we build an interactive platform (solo-LTR database) via integrating nearly complete genomic solo-LTR and LTR-RTs from 18 legume species., which enables the zero-code exploration of centromere-enriched repeats, serving as a crucial resource for designing synthetic centromeres. Together, our study reveals the Fabaceae centromere diversity, and provides mechanistic insights into the evolution of centromeres in legumes and a broad range of plant lineages.