MicroRNAs (miRNAs) are short noncoding RNAs of around 22 nucleotides in length, which help to shape the expression of most mRNAs. Perturbation of miRNA expression has revealed a variety of defects in development, cell specification, physiology and behavior. This thesis focuses on two topics of miRNA: identification of structural features that influence miRNA biogenesis (Paper I) and application of taxonomical marker miRNAs to resolve organismal origin of samples (Paper II and III).

The current model of miRNA hairpin biogenesis has limited information content and appears to be incomplete. In paper I, we apply a novel high-throughput screening method to profile the optimal structure of miRNA hairpins for efficient and precise miRNA biogenesis. The optimal structure consists of tight and loose local structures across the hairpin, which reflects the constraints of biogenesis proteins. We find that miRNA hairpins with stable lower basal stem are more efficiently processed and have a higher expression level in tissues of 20 animal species. We address that the structural features - which have been largely neglected in the current model - are in fact as important as the well-known sequence motifs.

New miRNAs are continuously added over evolutionary time and are rarely secondarily lost, making them ideal taxonomical markers. In paper II, we demonstrate as a proof-of-principle that miRNAs can be used to trace biological sample back to the lineage or even species of origin. Based on the marker miRNAs, we develop miRTrace, the first software to accurately trace miRNA sequences back to their taxonomical origin. The method can sensitively identify the origin of single cells and detect parasitic nematode RNA in mammalian host blood sample. In paper III, we apply miRNA tracing to address a controversial question about the origin of the exogenous plant miRNAs (xenomiRs) found in human samples, and which have been proposed to regulate human gene expression. Our computational and experimental results provide evidence that xenomiRs are derived from technical artifacts rather than dietary intake.