AUTHOR=Reczko Martin , Maragkakis Manolis , Alexiou Panagiotis , Papadopoulos Giorgio L., Hatzigeorgiou Artemis G. TITLE=Accurate microRNA target prediction using detailed binding site accessibility and machine learning on proteomics data JOURNAL=Frontiers in Genetics VOLUME=2 YEAR=2012 URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2011.00103 DOI=10.3389/fgene.2011.00103 ISSN=1664-8021 ABSTRACT=

MicroRNAs (miRNAs) are a class of small regulatory genes regulating gene expression by targeting messenger RNA. Though computational methods for miRNA target prediction are the prevailing means to analyze their function, they still miss a large fraction of the targeted genes and additionally predict a large number of false positives. Here we introduce a novel algorithm called DIANA-microT-ANN which combines multiple novel target site features through an artificial neural network (ANN) and is trained using recently published high-throughput data measuring the change of protein levels after miRNA overexpression, providing positive and negative targeting examples. The features characterizing each miRNA recognition element include binding structure, conservation level, and a specific profile of structural accessibility. The ANN is trained to integrate the features of each recognition element along the 3′untranslated region into a targeting score, reproducing the relative repression fold change of the protein. Tested on two different sets the algorithm outperforms other widely used algorithms and also predicts a significant number of unique and reliable targets not predicted by the other methods. For 542 human miRNAs DIANA-microT-ANN predicts 120000 targets not provided by TargetScan 5.0. The algorithm is freely available at http://microrna.gr/microT-ANN.