Human Genome Encodes Multiple Novel miRNAs

Human cells generate a vast complexity of non-coding RNAs, the so called “RNA dark matter”, which includes small RNA transcriptome represented by hundreds of thousands of entities which outnumber members of known classes of small RNAs by orders of magnitude. Biogenesis, biological relevance, and mechanisms of action of most of these transcripts remain unknown and many of them are assumed to represent degradation products of known genes. In this study, we aimed to functionally characterize human sRNA transcriptome by attempting to answer the following question — can a significant number of novel sRNAs correspond to novel members of known classes of sRNA, specifically, miRNAs? We show that at the very least, 2726 novel miRNAs exist in just one human cell line. Furthermore, potentially thousands other novel members of this class may exist. Strikingly, many novel miRNAs are derived from exons of protein-coding genes. These results suggest that mammalian novel sRNA transcriptome harbors multitude of novel functional transcripts, at least some of which belong to known classes of sRNAs. Overall design: We developed a miRNA discovery pipeline that detects novel miRNAs that combine the features of being generated in vivo using the Drosha/DGCR8 biogenesis pathway with correct RNA secondary structure. We further validated the pipeline by detecting in vivo signatures of Dicer cleavage. Using this approach, we discovered thousands of novel miRNAs in just one human cell line and suggest that many others exist.