Etre ou ne pas être un microRNA: l'existence précaire d'un régulateur capital

Daniel Gautheret (Univ. Paris XI)

Que faut-il pour qu'un microRNA fonctionnel s'exprime? Peu de choses en fait. Les enzymes qui forment les miRNA reconnaissent des structures tige-boucle très fréquentes dans tous les génomes et qui se retrouvent dans de nombreux transcrits. Je montrerai avec quelle facilité les microRNA apparaissent et disparaissent au cours de l'évolution, puis j'expliquerai comment on peut exploiter les méthodes de comparaison de structure pour tenter de mieux cerner l'essence d'un microRNA. Je présenterai enfin notre serveur d'évaluation des candidats microRNA "miReval".

Caractérisation des microARN chez un insecte phytophage : le puceron du pois Acyrthosiphon pisum

Stéphanie Jaubert et Denis Tagu (INRA, Rennes)

MicroRNA Target Search and Genome Wide Detection of Small RNAs

Stephan Ossowski (Max Planck Institute for Developmental Biology, Tübingen)

Small RNAs of 19-24 nucleotides are mediators of gene silencing and transcriptional regulation that bind to target transcripts through complementary base pairing. MicroRNAs, one class of silencing RNAs, originate from a characteristic hairpin-containing transcript and mediate cleavage or translational inhibition of target mRNAs. Vectors that contain a miRNA hairpin precursor are recognized as second-generation RNAi silencing vectors. Their sequence can be modified such that artificial miRNAs (amiRNAs) targeting a defined set of genes are produced in vivo. AmiRNAs have several advantages compared to other silencing methods, most importantly they can silence several genes and whole gene families at once, have little off-target effects and can be induced in a temporally and spatially controlled fashion.

We have developed a miRNA target search tool based on short sequence alignment, RNA-RNA folding and multiple filters for known features of miRNA target site recognition. MiRNA sequences are optimally aligned against the reference genome using enhanced suffix arrays (http://www.vmatch.de) allowing for up to five mismatches and/or gaps. RNA-RNA hybridization energy is calculated using the Vienna RNA package. The TargetSearch algorithm was subsequently used to calculate highly efficient amiRNAs targeting one or more selected genes while avoiding any off-target effects. AmiRNAs can be calculated for all (at least partially) sequenced plant genomes using our web-based tool WMD. Several studies have shown that amiRNAs produced by WMD work in many plant species, from unicellular algae (C. reinhardtii), to mosses (P. patents), monocots (O. sativa) and dicots (A. thaliana).

We have recently used our experience in short sequence alignment to develop new methods for small RNA detection with Next Generation Sequencing (NGS) using the Illumina GA2, often called sRNA-Seq. With ~100 Million reads per week and a relatively high error rate NGS poses new challenges for short sequence alignments and subsequent consensus analysis. Our sequencing pipeline SHORE incorporates all necessary steps from quality filtering, fast whole genome alignment, sRNA detection and visualization as well as statistical comparison of multiple sRNA samples. To enable the alignment of millions of sRNAs within a few hours we have developed a short read alignment tool called GenomeMapper. Sequences are aligned against the reference genome allowing for mismatches and gaps, considering both unique and repetitive hits. Regions of consecutive read coverage correspond to small RNA clusters or mature miRNAs. Differences in abundance of sRNAs between multiple samples are identified using a linear mixed model also correcting for background noise and run specific effects.
WMD is available at http://wmd2.weigelworld.org. SHORE and GenomeMapper will soon be available at http://www.1001genomes.org.

The contributors to this work are S. Ossowski, R. Schwab, N. Warthmann, K. Schneeberger, L. Smith, J. Hagmann, F. Ott, A. Schützenmeister, C. Lanz, D. Weigel