Stochastic alternative splicing is prevalent in mungbean (Vigna radiata)

Abstract

Alternative splicing (AS) can produce multiple mature mRNAs from the same primary transcript, thereby generating diverse proteins and phenotypes from the same gene. To assess the prevalence of AS in mungbean (Vigna radiata), we analysed whole-genome RNA sequencing data from root, leaf, flower and pod tissues and found that at least 37.9% of mungbean genes are subjected to AS. The number of AS transcripts exhibited a strong correlation with exon number and thus resembled a uniform probabilistic event rather than a specific regulatory function. The proportion of frameshift splicing was close to the expected frequency of random splicing. However, alternative donor and acceptor AS events tended to occur at multiples of three nucleotides (i.e. the codon length) from the main splice site. Genes with high exon number and expression level, which should have the most AS if splicing is purely stochastic, exhibited less AS, implying the existence of negative selection against excessive random AS. Functional AS is probably rare: a large proportion of AS isoforms exist at very low copy per cell on average or are expressed at much lower levels than default transcripts. Conserved AS was only detected in 629 genes (2.8% of all genes in the genome) when compared to Vigna angularis, and in 16 genes in more distant species like soya bean. These observations highlight the challenges of finding and cataloguing candidates for experimentally proven AS isoforms in a crop genome.

Keywords: RNA sequencing; alternative splicing; evolutionary conservation; mungbean (Vigna radiata); stochastic process.

Figure 1

Types and chromosomal distribution of AS in four mungbean tissues. From outer to inner rings: (a) size of chromosome (in megabases); (b) histogram of AS number across chromosomes in root, (c) leaf, (d) flower and (e) pod tissues. (f) Proportions of each type of AS across the four tissues, as classified by ASTALAVISTA.

Figure 2

Correlation of mean AS number with number of exons in a gene (a) and gene expression level estimated from alignment coverage (b).

Figure 3

Comparison of the average number of AS events per exon, calculated by dividing the number of AS events in a gene with the number of exons in that gene, for genes containing different numbers of exons (a) and genes expressed at different levels (b), as determined by alignment coverage per base.

Figure 4

Number of AS isoforms, according to distance from the regular splice site, for alternative donor and alternative acceptor isoform types.

Figure 5

Proportion of DNA bases surrounding alternative splice types compared to the regular splice sites. AS isoforms with FPKM >10 were categorized as having high concentration (H), while those with FPKM <10 were grouped into the low concentration group (L).

Figure 6

Gene ontology enrichment of genes with AS isoforms that are conserved in mungbean and adzuki bean.