In plants, Dicer-like 1 (DCL1)-mediated two-step cleavages are essential for the processing of microRNA (miRNA) gene products. Interestingly, DCL1 has been indicated to be involved in the production of many small RNAs (sRNAs) that cannot be classified as canonical miRNAs. However, genomic and functional information on the non-miRNA, DCL1-dependent sRNAs is still limited. Here, we propose a secondary structure-based approach for identification of the precursors containing novel DCL1-dependent sRNA loci. To demonstrate the utility of the workflow: first, 5898 DCL1-dependent sRNAs of 20-24 nucleotides were identified from the sRNA high-throughput sequencing data sets prepared from rice DCL1 RNA interference transgenic lines. Those perfectly mapped to the rice pre-miRNAs (precursor microRNAs) were removed. The remaining 5795 sRNAs were then mapped onto the rice genome, obtaining 30 902 perfectly matched loci belonging to 2310 sRNAs. A total of 4631 clusters of sRNA loci were defined for secondary structure prediction by using RNAfold. The prediction results generated by two algorithms, namely MFE (minimum free energy) and centroid, were manually compared to identify the conserved long-stem structures containing DCL1-dependent sRNA loci. For the purpose of a case study, a portion of the prediction results was screened manually. As a result, 60 clusters displayed great potential for forming featured long-stem structures for the generation of DCL1-dependent sRNAs. Together, the results indicate that the proposed workflow is applicable for the identification of novel DCL1-dependent sRNA loci on plant genomes.
Keywords: Cluster; Dicer-like 1 (DCL1); high-throughput sequencing; microRNA (miRNA); precursor microRNAs (pre-miRNAs); secondary structure; small RNA (sRNA)..