1. Structure and sequence features determining miRNA biogenesis

The human transcriptome contains more than 100,000 hairpin RNA structures, more than half of which are located in mRNAs. These hairpins are potential entry points into miRNA biogenesis, and thus constitute cross-roads for nuclear transcripts. They can either be cleaved into regulatory miRNAs or they can avoid cleavage and function as full-length transcripts, for instance through cytoplasmic transport and translation. Currently little is understood of the features that determine this critical decision, and attempts to elucidate them are hampered by the lack of genomic methods in the field of miRNA biogenesis. We are currently developing the first method to test hairpin cleavage transcriptome-wide. We will use this method to profile the vast majority of human hairpin structures, and will apply machine-learning approaches to extract the structure and sequence features which license or block the miRNA biogenesis.


Transcripts containing RNA hairpins are either cleaved to miRNAs or function as full-length transcripts, for instance by cytoplasmic export and translation. It is not known what determines this critical decision.

2. miRNA functions in single cells

It is well established that miRNAs can strongly down-regulate their mRNA targets during dynamic processes such as development. However, the functions of miRNAs in steady-state conditions are not well understood, although there is emerging evidence that they can buffer oscillations of their target mRNA expression, thus stabilizing transcriptional profiles. Since many oscillations are not synchronized between cells, such information is lost when thousands of cells are profiled at the same time. We will study the expression of miRNA targets in single cells using state-of-the-art next generation sequencing protocols. Comparing the behavior of the target mRNAs in wild-type cells and in mutant cells void of miRNAs, we will estimate and model the impact of miRNAs on single-cell transcriptomes.