Assembly of complete genomes can reveal functional genetic elements missing from draft sequences. Here we present the near-complete telomere-to-telomere and contiguous genome of the cotton species Gossypium raimondii. Our assembly identified gaps and misoriented or misassembled regions in previous assemblies and produced 13 centromeres, with 25 chromosomal ends having telomeres. In contrast to satellite-rich Arabidopsis and rice centromeres, cotton centromeres lack phased CENH3 nucleosome positioning patterns and probably evolved by invasion from long terminal repeat retrotransposons. In-depth expression profiling of transposable elements revealed a previously unannotated DNA transposon (MuTC01) that interacts with miR2947 to produce trans-acting small interfering RNAs (siRNAs), one of which targets the newly evolved LEC2 (LEC2b) to produce phased siRNAs. Systematic genome editing experiments revealed that this tripartite module, miR2947–MuTC01–LEC2b, controls the morphogenesis of complex folded embryos characteristic of Gossypium and its close relatives in the cotton tribe. Our study reveals a trans-acting siRNA-based tripartite regulatory pathway for embryo development in higher plants.

完整基因组的组装可以揭示草图序列中缺失的功能遗传元件。在这里，我们展示了雷蒙德棉棉种近乎完整的端粒到端粒和连续基因组。我们的组装鉴定了先前组装中的间隙和错误定向或错误组装的区域，并产生了 13 个着丝粒，其中 25 个染色体末端具有端粒。与富含卫星的拟南芥和水稻着丝粒相比，棉花着丝粒缺乏阶段性CENH3核小体定位模式，并且可能是通过长末端重复逆转录转座子的入侵而进化的。转座元件的深入表达谱揭示了一个先前未注释的DNA转座子（MuTC01），它与miR2947相互作用产生反式作用小干扰RNA（siRNA），其中一个靶向新进化的LEC2 （ LEC2b ）以产生定相siRNA。系统的基因组编辑实验表明，这个三部分模块，miR2947-MuTC01- LEC2b ，控制着棉族及其近亲的复杂折叠胚胎特征的形态发生。我们的研究揭示了高等植物胚胎发育的基于反式作用siRNA的三方调控途径。