Genome evolution and initial breeding of the Triticeae grass <i>Leymus chinensis</i> dominating the Eurasian Steppe

<i>Leymus chinensis</i>, a dominant perennial grass in the Eurasian Steppe, is well known for its remarkable adaptability and forage quality. Hardly any breeding has been done on the grass, limiting its potential in ecological restoration and forage productivity. To enable genetic improvement of the untapped, important species, we obtained a 7.85-Gb high-quality genome of <i>L. chinensis</i> with a particularly long contig N50 (318.49 Mb). Its allotetraploid genome is estimated to originate 5.29 million years ago (MYA) from a cross between the Ns-subgenome relating to <i>Psathyrostachys</i> and the unknown Xm-subgenome. Multiple bursts of transposons during 0.433-1.842 MYA after genome allopolyploidization, which involved predominantly the Tekay and Angela of LTR retrotransposons, contributed to its genome expansion and complexity. With the genome resource available, we successfully developed a genetic transformation system as well as the gene-editing pipeline in <i>L. chinensis</i>. We knocked out the monocot-specific miR528 using CRISPR/Cas9, resulting in the improvement of yield-related traits with increases in the tiller number and growth rate. Our research provides valuable genomic resources for Triticeae evolutionary studies and presents a conceptual framework illustrating the utilization of genomic information and genome editing to accelerate the improvement of wild <i>L. chinensis</i> with features such as polyploidization and self-incompatibility.