CentriVision: An integrated platform for multiscale centromere analysis in plants.

Centromeres are essential for accurate chromosome segregation and genome stability; with the advent of telomere-to-telomere genome assemblies, they have become central targets of genome-wide studies. Here, we present CentriVision, a modular bioinformatics platform that integrates candidate centromere identification, structural similarity assessment, DNA repeat unit decomposition, and a framework for exploring potential relationships between single-nucleotide conservation and functional features. CentriVision provides a comprehensive suite of analytical tools, including edit-distance dot plots, intra-segment heatmaps, kilobase-scale mini-dot plots, repeat monomer scanning with conserved-site visualization, and satellite DNA expansion-divergence estimation, all of which can be seamlessly integrated with CENH3 chromatin immunoprecipitation sequencing (ChIP-seq) data. When applied to representative plant species, CentriVision achieved high predictive accuracy and revealed diverse organizational patterns. Arabidopsis thaliana centromeres are primarily composed of 178-188-bp repeats interspersed with rarer ∼502-bp variants that exhibit pronounced sequence conservation but only background CENH3-ChIP signal, suggesting that these elements represent pre-centromeric sequences overlooked in earlier studies. Oryza sativa contains two dominant classes of centromeric repeats rather than the single class previously reported. In contrast, Zea mays exhibits strongly biased expansion toward the evolution of a single dominant repeat unit, reflecting a distinct evolutionary strategy of centromere reconstruction, whereas Papaver setigerum displays a notable three-layered nested repeat structure. Integration of repeat sequence divergence with CENH3 binding further revealed lineage-specific evolutionary trajectories of centromere specification. Collectively, these findings advance our understanding of centromere structure and function. CentriVision offers a reproducible, scalable, and user-friendly framework that quantitatively links repeat evolution, structural variation, and functional epigenomics, providing new insights into the architecture and diversification of plant centromeres.