A Genome-wide Atlas of MicroRNA Primary Transcript Structure

MicroRNA (miRNA) expression is dynamically regulated during development, across tissues, and in various human diseases. Nevertheless, a major bottleneck in the elucidation of mechanisms that control miRNA abundance is the currently incomplete annotation of primary miRNA (pri-miRNA) gene structures. While a subset of miRNAs are hosted in protein-coding genes, the majority of pri-miRNAs are transcribed as poorly-characterized noncoding transcripts. Due to the efficiency of DROSHA processing, the abundance of pri-miRNAs is very low at steady-state. Therefore, elucidation of pri-miRNA structure has remained a significant challenge. To address this problem, we developed an experimental and computational approach that allows rapid transcriptome-wide mapping of pri-miRNA structures. By performing deep RNA-seq in cells expressing a dominant-negative DROSHA mutant protein, we demonstrated dramatic enrichment of intact pri-miRNAs, resulting in much greater coverage of these transcripts compared to standard RNA-seq. We also evaluated multiple assembly algorithms, ultimately demonstrating that StringTie, a software package that we recently developed, outperforms other existing assembly tools for this application. We applied our novel pri-miRNA annotation strategy to a panel of human and mouse cell lines of diverse origins, thereby significantly improving the existing annotation of conserved mammalian miRNA genes. Based on these new assemblies, we uncovered unanticipated features and new potential regulatory mechanisms, including unexpected links between pri-miRNAs and distant protein coding genes, alternative pri-miRNA splicing, and alternative promoter usage that can produce transcripts carrying subsets of miRNAs encoded by polycistronic clusters. These results provide a valuable resource for the study of mammalian miRNA regulation.