Identification of functional elements involved in RNA splicing, domestication, and agronomic performance in mungbean genome

Abstract

Mungbean (Vigna radiata) is an important Asian pulse crop with annual cultivation area of 6 million hectares. The completion of mungbean reference genome provided a major boost for mungbean genomic research. Annotation of the reference genome data was improved by evaluating the prevalence of alternative splicing (AS) in the genome. At least 37.9% of mungbean genes undergo AS but there are indications that AS in mungbean is predominantly stochastic. A large proportion of AS isoforms exists at very low copy, or expressed at much lower level than the default transcript. Conservation in closely related species is also rare, with only 2.8% of genes share conserved AS between mungbean and adzuki bean, and only 16 soybean genes share conserved AS with mungbean. Using genotyping by sequencing (GBS) on 276 cultivated and wild mungbean accessions from around the world, it was found that nucleotide diversity among cultivated accessions decreases to 30% of the level found in wild accessions. Linkage disequilibrium also decays at longer interval in cultivated mungbean, where LD blocks are on average 4.6 times longer than wild accessions. Wild and cultivated accessions are clearly separated in phylogenetic and population structure analysis, and there are some correlations between geographic origin and subpopulation membership. Several loci were identified as possible candidate for regions that underwent positive selection during mungbean domestication. The genes in those intervals are enriched with genes associated with growth and reproductive traits. SNP data obtained from GBS were also used for genome-wide association study (GWAS) to dissect the genetic background of several agronomic traits such as flowering time, maturity time, pod formation time, seed weight, number of seeds per pod, peak harvest, cumulative weekly harvest, final yield, and synchronicity. Phenotyping was performed on 222 cultivated accessions and using two different association methods at least 79 markers were found to be significantly associated with the traits at p-value <0.0001. Some of the genes that intersect with significant markers share homology with soybean genes and QTL that could explain their role in trait formation, which makes them attractive candidates for follow up studies. The data can be used as a basis for mapping studies or parental selection in mungbean breeding programs.