Genetic diversity and evolution of rice centromeres

Understanding the driving force of centromere dynamics is crucial for deciphering the complexity of eukaryotic evolution and speciation. Here we assembled 67 rice genomes from the Oryza AA group and analyzed >800 nearly complete centromeres. Through de novo annotation of centromeric satellite CEN155 sequences and employing a progressive compression strategy, we quantified the local homogenization and multilayer structures of rice satellite arrays. Our results indicate that genetic innovations in rice centromeres primarily arise from structural variations and centrophilic retrotransposon insertions. The single-base substitution rate in rice centromeres appears to be lower relative to that in chromosome arms. Comparisons of CEN155 arrays, retrotransposons and functional centromeres highlight their dynamic but correlated interplay. Contrary to the KARMA model for Arabidopsis centromere evolution, we propose a hypothesis that retrotransposon invasion probably contributes to the decline of progenitor centromeric satellite arrays and promotes centromere repositioning, as evidenced by extended CENH3 chromatin immunoprecipitation sequencing enrichment beyond the native satellite arrays.