Computational prediction of candidate miRNAs and their targets from medicago truncatula non-protein-coding transcripts

Identification and analysis of miRNAs enhances our understanding of the important roles that small RNAs play in complex regulatory networks. It is often difficult to perform large-scale validation of miRNA expression that is predicted from genomic regions. Expressed transcripts provide an alternative resource to facilitate identification of miRNAs and their targets. We developed a computational pipeline to scan for miRNA genes from polyadenylated transcripts that were associated with limited protein coding potentials, corresponding to the intergenic regions of Medicago truncatula genomic sequences. Each predicted miRNA was required to have a near perfect match with target genes. We also searched for miRNA conservation in other plant species, clustered highly similar miRNAs, and provided a functional classification of target genes. The data is represented in the Medicago-MIRATdb (MiRNA And Target gene Data Base). The database provides detailed information on the sequences of the predicted miRNAs, their precursors, and potential target genes. It also details sequence source information such as the EST library, tissue category, and the number of EST clones. Information regarding miRNA conservation in other species, functional classification of target genes, and clusters of similar miRNAs is also provided. The web interface to the database provides researchers with the ability to narrow their search for miRNAs and target genes of interest by using a variety of filters. To our knowledge, this work represents the first systematic effort to identify candidate miRNAs and targets in the model legume M. truncatula.