miRNAfe: a tool for feature extraction in pre-miRNA prediction

C. YONES, G. STEGMAYER, L. KAMENETZKY, D.H. MILONE

Conferencia; VI Argentinian Conference on Bioinformatics and Computational Biology; 2015

MicroRNAs (miRNA) are a group of short (~22 nucleotides) non-coding RNA which can play important roles ingene regulation by targeting mRNAs for cleavage or translational repression [1]. Precursors of miRNA (premiRNA)are characterized by their hairpin structure. However, a large amount of similar sequences can be foldedinto this kind of structure in many genomes.In order to predict miRNAs, the first step is to extract features from sequences of a complete genome to train abinary classifier. After that, the classifier is used to predict which sequences from a genome are more likely to bea pre-miRNA. The feature extraction step is very important for this whole process, in order to achieve high ratesof true positives predictions. The amount of features that have been proposed in the last years is very large andthere are many different tools that partially achieve this task. They are coded in different programming languagesand have different access modes (web, command line, etc.). Besides, several tools are proprietary software andthe source code is not even available.