A Gapless T2T Genome and Population-Scale Resequencing of Adzuki Bean Reveal Centromere Evolution and Accelerate Molecular Breeding
Changyou Liu, Jianxiang Peng, Huilong Qi, Shen Wang, Huiying Shi, Xiaochang Xu, Yirong Yang, Tianmeng Wang, Yan Wang, Qian Yu, Huizhe Zhang, Gaoli Hu, Zhixiao Zhang, Jifeng Liang, Meifang Lan, Yingchao Shen, Yongchao Jin, Linlin Gu, Tao Liu, Gaoling Luo, Yanhua Chen, Xiaoyan Zhang, Fengjing Song, Lizhen Zheng, Shutong Li, Jijun Zhang, Juechen Long, Guofang Wu, Ning Xu, Kunpeng Deng, Changyan Liu, Wenzhao Xie, Jing Tian, Bingbing Wang, Xiyin Wang, Baojie FanAbstract
The adzuki bean (Vigna angularis) is an economically and nutritionally significant legume species. Using a multi-strategy sequencing approach, a gapless, telomere-to-telomere (T2T) genome assembly of an elite cultivar, Jihong 16 (JH16), was generated. The genome size was 514.85 Mb, with a Contig N50 of 48.72 Mb, 3x longer than previous published genome. The centromeres of adzuki bean were revealed for the first time and found to be composed of ~ 9-bp minisatellites — a distinct departure from the much longer satellites typically found in other plants. Low sequence conservation between the centromeres of adzuki bean and mung bean further highlights the highly active evolutionary dynamics of centromeric sequences in Vigna species. Whole-genome resequencing of 706 adzuki bean accessions was performed and three major phylogenetic groups were identified. Among these, the Southern China group exhibited the highest genetic diversity and the most rapid linkage disequilibrium (LD) decay. A set of 112 core accessions was selected to represent the global germplasm, and 12 barcode SNPs were identified for DNA fingerprinting. Leveraging high-quality phenotypic data for 39 agronomic traits collected across seven locations over two years, a genome-wide association study (GWAS) identified 9613 significant marker-trait associations (MTAs) and thousands of candidate genes, including 40 MTAs located in the centromere regions. Several genes were experimentally validated, including the flowering-time gene Vigan07G000670, overexpression of which in Arabidopsis led to early flowering phenotype and altered FT/FLC expression. Finally, genomic selection (GS) evaluations demonstrated that even a small set (500 ~ 1000) of trait-associated markers achieved high prediction accuracy (>0.8) for key agronomy traits such as seed weight, flowering date and plant height. The genomic resources and insights generated in this study provide a robust foundation for genetic improvement of adzuki bean and related legume crops.