Computational Prediction of MicroRNA Target Genes, Target Prediction Databases, and Web Resources

MicroRNA (miRNA) mediated silencing and repression of mRNA molecules requires complementary base pairing between the "seed" region of the miRNA and the "seed match" region of target mRNAs. While this mechanism is fairly well understood, accurate prediction of valid miRNA targets remains challenging due to factors such as imperfect sequence specificity, target site availability, and the thermodynamic stability of the mRNA structure itself. As knowledge of what genes are being targeted by each miRNA is arguably the most important facet of miRNA biology, many approaches have been developed to address the need for reliable prediction and ranking of putative targets, with most using a combination of various strategies such as evolutionary conservation, statistical inference, and distinct features of the target sequences themselves. This chapter reviews the pros and cons of a number of different prediction algorithms, showcases some databases that store experimentally validated miRNA targets, and also provides a case study that profiles some of the potential microRNA-mRNA interactions predicted by each methodology for various human genes.