Refining microRNA target predictions: Sorting the wheat from the chaff

https://doi.org/10.1016/j.bbrc.2014.01.181Get rights and content

Highlights

  • The principles underlying most commonly used microRNA target prediction algorithms.

  • An evaluation of the experimental data that supports each computational approach.

  • An evaluation of microRNA target databases, their strengths and weaknesses.

Abstract

microRNAs are short RNAs that reduce gene expression by binding to their targets. The accurate prediction of microRNA targets is essential to understanding the function of microRNAs. Computational predictions indicate that all human genes may be regulated by microRNAs, with each microRNA possibly targeting thousands of genes. Here we discuss computational methods for identifying mammalian microRNA targets and refining them for further experimental validation. We describe microRNA target prediction resources and procedures and how they integrate with various types of experimental techniques that aim to validate them or further explore their function. We also provide a list of target prediction databases and explain how these are curated.

Introduction

microRNAs (miRNAs) are short, ∼22 nucleotide long RNAs that reduce gene expression, usually by binding to the 3′ untranslated region of target mRNAs. miRNAs guide a protein complex called RNA-induced silencing complex (RISC) to specific mRNA target sites called microRNA responsive elements (mREs). miRNAs were first discovered in 1993 during an analysis of larval developmental timing in the worm Caenorhabditis Elegans, where a 22-nucleotide RNA regulated protein abundance of LIN-14 [1]. First regarded as a Nematode-specific RNA family, it was only in 2000 that another microRNA, let-7, was characterized and identified in other species. In 2002, Eric Lai compared the sequences of 11 microRNAs to the K box and Brd Box motifs that were known to mediate post-transcriptional regulation in Drosophila. He demonstrated that the first eight nucleotides, now called the seed region, of miRNAs, were perfectly complementary to these motifs and concluded that this complementarity may be essential in post-transcriptional regulation by microRNAs [2]. This simple bioinformatics analysis established one of the strongest predictive features used in target prediction to date and was the basis for numerous algorithms that enabled the explosion of miRNA functional characterization.

To date over 1000 miRNAs have been identified in Humans, hundreds of which are associated with major biological processes including cell proliferation and differentiation, development and disease. As such miRNAs are arguably one of the most important classes of functional RNAs. However, the rules governing miRNA target recognition are not fully understood and may vary for each miRNA–mRNA pair. Computational approaches that can test various models of miRNA binding and predict target sites are therefore essential to understanding the function of microRNAs.

In this review, we will outline the major concepts underlying the most popular target prediction algorithms, discuss the context in which these algorithms are optimal and provide simple techniques to refine them. A list of readily available miRNA target prediction algorithms and databases is given in Table 1.

Section snippets

Materials and methods

All the techniques and approaches discussed here are available online.

Results

An ouroboric relation has always existed between the computational prediction of miRNA targets and their experimental validation. Experimental approaches are often designed and interpreted by using computational predictions and these in turn have been created and refined based on experimental output. It is therefore difficult to obtain an independent experimental dataset to compare the efficiency of target prediction algorithms and until recently, experiments were not designed to discover novel

Discussion

Experimentalists using miRNA target prediction software for the first time will invariably observe two apparent shortfalls. The first is that target prediction algorithms produce large lists of candidates, many of which cannot currently be validated experimentally. These false positives are generally attributed to a poor modelling of the miRNA–mRNA interaction. This may not always be the case. Recent studies have demonstrated that miRNAs, their guide, and their effector proteins are tightly

References (25)

  • S.W. Chi et al.

    Argonaute HITS-CLIP decodes microRNA-mRNA interaction maps

    Nature

    (2009)
  • R.C. Friedman et al.

    Most mammalian mRNAs are conserved targets of microRNAs

    Genome Res.

    (2009)
  • Cited by (0)

    View full text