A reality check for alignments and trees

Making multiple sequence alignments is one of the more commonplace procedures in modern biology. Multiple alignments are typically generated by feeding sequences into the alignment program from the N-terminus to the C-terminus. Recent results show that if the same sequences are processed from the C- to the N-terminus, a different alignment is often obtained. Because phylogenetic trees are built from alignments, the resulting trees can also differ. The new findings highlight sequence alignment as a crucial step in molecular evolutionary studies and provide straightforward measures to assess alignment reliability.     

miRTar Hunter: A prediction system for identifying human microRNA target sites

MicroRNAs (miRNAs) are important regulators of gene expression and play crucial roles in many biological processes including apoptosis, differentiation, development, and tumorigenesis. Recent estimates suggest that more than 50% of human protein coding genes may be regulated by miRNAs and that each miRNA may bind to 300–400 target genes. Approximately 1,000 human miRNAs have been identified so far with each having up to hundreds of unique target mRNAs. However, the targets for a majority of these miRNAs have not been identified due to the lack of large-scale experimental detection techniques. Experimental detection of miRNA target sites is a costly and time-consuming process, even though identification of miRNA targets is critical to unraveling their functions in various biological processes. To identify miRNA targets, we developed miRTar Hunter, a novel computational approach for predicting target sites regardless of the presence or absence of a seed match or evolutionary sequence conservation. Our approach is based on a dynamic programming algorithm that incorporates more sequence-specific features and reflects the properties of various types of target sites that determine diverse aspects of complementarities between miRNAs and their targets. We evaluated the performance of our algorithm on 532 known human miRNA:target pairs and 59 experimentally-verified negative miRNA:target pairs, and also compared our method with three popular programs for 481 miRNA:target pairs. miRTar Hunter outperformed three popular existing algorithms in terms of recall and precision, indicating that our unique scheme to quantify the determinants of complementary sites is effective at detecting miRNA targets. miRTar Hunter is now available at http://203.230.194.162/~kbkim.     

Improving performance of mammalian microRNA target prediction

Background  

MicroRNAs (miRNAs) are single-stranded non-coding RNAs known to regulate a wide range of cellular processes by silencing the gene expression at the protein and/or mRNA levels. Computational prediction of miRNA targets is essential for elucidating the detailed functions of miRNA. However, the prediction specificity and sensitivity of the existing algorithms are still poor to generate meaningful, workable hypotheses for subsequent experimental testing. Constructing a richer and more reliable training data set and developing an algorithm that properly exploits this data set would be the key to improve the performance current prediction algorithms.     

HOCTAR database: a unique resource for microRNA target prediction

An Antarctic strain (NJ-7) of Chlorella vulgaris possesses the same 18S rRNA sequence as that of a temperate strain (UTEX259), but shows significantly higher freezing tolerance than the latter. Suppression subtractive hybridization (SSH) was performed to identify genes of intensified expression in NJ-7 relative to UTEX259. Among the genes identified, Ccor1 and Ccor2, co-organized in the same gene cluster Ccor1-Ccor2-Ccor1-Ccor2, showed much higher expression levels in NJ-7 than in UTEX259 at both 20°C and 4°C. As detected by Northern blot and Western blot analyses, the two genes were cold-inducible in NJ-7 but almost not expressed in UTEX259. Their encoded products are predicted to share 55.7% identity to each other and possess physicochemical characteristics similar to that of late embryogenesis abundant (LEA) proteins in plants. The purified recombinant Ccor1 and Ccor2 showed high heat-stability and could act as cryoprotectants to lactate dehydrogenase in vitro. Based on their expression patterns and protein characteristics, we propose that Ccor1 and Ccor2 are two novel LEA proteins and are related to the greatly enhanced freezing tolerance in the Antarctic strain.     

Efficient use of accessibility in microRNA target prediction

Considering accessibility of the 3′UTR is believed to increase the precision of microRNA target predictions. We show that, contrary to common belief, ranking by the hybridization energy or by the sum of the opening and hybridization energies, used in currently available algorithms, is not an efficient way to rank predictions. Instead, we describe an algorithm which also considers only the accessible binding sites but which ranks predictions according to over-representation. When compared with experimentally validated and refuted targets in the fruit fly and human, our algorithm shows a remarkable improvement in precision while significantly reducing the computational cost in comparison with other free energy based methods. In the human genome, our algorithm has at least twice higher precision than other methods with their default parameters. In the fruit fly, we find five times more validated targets among the top 500 predictions than other methods with their default parameters. Furthermore, using a common statistical framework we demonstrate explicitly the advantages of using the canonical ensemble instead of using the minimum free energy structure alone. We also find that ‘naïve’ global folding sometimes outperforms the local folding approach.     

TargetSpy: a supervised machine learning approach for microRNA target prediction

Background  

Virtually all currently available microRNA target site prediction algorithms require the presence of a (conserved) seed match to the 5' end of the microRNA. Recently however, it has been shown that this requirement might be too stringent, leading to a substantial number of missed target sites.     

A non-canonical plant microRNA target site

Plant microRNAs (miRNAs) typically form near-perfect duplexes with their targets and mediate mRNA cleavage. Here, we describe an unconventional miRNA target of miR398 in Arabidopsis, an mRNA encoding the blue copper-binding protein (BCBP). BCBP mRNA carries an miR398 complementary site in its 5′-untranslated region (UTR) with a bulge of six nucleotides opposite to the 5′ region of the miRNA. Despite the disruption of a target site region thought to be especially critical for function, BCBP mRNAs are cleaved by ARGONAUTE1 between nucleotides 10th and 11th, opposite to the miRNA, like conventional plant target sites. Levels of BCBP mRNAs are inversely correlated to levels of miR398 in mutants lacking the miRNA, or transgenic plants overexpressing it. Introducing two mutations that disrupt the miRNA complementarity around the cleavage site renders the target cleavage-resistant. The BCBP site functions outside of the context of the BCBP mRNA and does not depend on 5′-UTR location. Reducing the bulge does not interfere with miR398-mediated regulation and completely removing it increases the efficiency of the slicing. Analysis of degradome data and target predictions revealed that the miR398-BCBP interaction seems to be rather unique. Nevertheless, our results imply that functional target sites with non-perfect pairings in the 5′ region of an ancient conserved miRNA exist in plants.     

Optimal use of conservation and accessibility filters in microRNA target prediction

It is generally accepted that filtering microRNA (miRNA) target predictions by conservation or by accessibility can reduce the false discovery rate. However, these two strategies are usually not exploited in a combined and flexible manner. Here, we introduce PACCMIT, a flexible method that filters miRNA binding sites by their conservation, accessibility, or both. The improvement in performance obtained with each of these three filters is demonstrated on the prediction of targets for both i) highly and ii) weakly conserved miRNAs, i.e., in two scenarios in which the miRNA-target interactions are subjected to different evolutionary pressures. We show that in the first scenario conservation is a better filter than accessibility (as both sensitivity and precision are higher among the top predictions) and that the combined filter improves performance of PACCMIT even further. In the second scenario, on the other hand, the accessibility filter performs better than both the conservation and combined filters, suggesting that the site conservation is not equally effective in rejecting false positive predictions for all miRNAs. Regarding the quality of the ranking criterion proposed by Robins and Press and used in PACCMIT, it is shown that top ranking interactions correspond to more downregulated proteins than do the lower ranking interactions. Comparison with several other target prediction algorithms shows that the ranking of predictions provided by PACCMIT is at least as good as the ranking generated by other conservation-based methods and considerably better than the energy-based ranking used in other accessibility-based methods.     

A functional assay for microRNA target identification and validation

MicroRNAs (miRNA) are a class of small RNA molecules that regulate numerous critical cellular processes and bind to partially complementary sequences resulting in down-regulation of their target genes. Due to the incomplete homology of the miRNA to its target site identification of miRNA target genes is difficult and currently based on computational algorithms predicting large numbers of potential targets for a given miRNA. To enable the identification of biologically relevant miRNA targets, we describe a novel functional assay based on a 3'-UTR-enriched library and a positive/negative selection strategy. As proof of principle we have used mir-130a and its validated target MAFB to test this strategy. Identification of MAFB and five additional targets and their subsequent confirmation as mir-130a targets by western blot analysis and knockdown experiments validates this strategy for the functional identification of miRNA targets.     

A framework for improving microRNA prediction in non-human genomes

The prediction of novel pre-microRNA (miRNA) from genomic sequence has received considerable attention recently. However, the majority of studies have focused on the human genome. Previous studies have demonstrated that sensitivity (correctly detecting true miRNA) is sustained when human-trained methods are applied to other species, however they have failed to report the dramatic drop in specificity (the ability to correctly reject non-miRNA sequences) in non-human genomes. Considering the ratio of true miRNA sequences to pseudo-miRNA sequences is on the order of 1:1000, such low specificity prevents the application of most existing tools to non-human genomes, as the number of false positives overwhelms the true predictions. We here introduce a framework (SMIRP) for creating species-specific miRNA prediction systems, leveraging sequence conservation and phylogenetic distance information. Substantial improvements in specificity and precision are obtained for four non-human test species when our framework is applied to three different prediction systems representing two types of classifiers (support vector machine and Random Forest), based on three different feature sets, with both human-specific and taxon-wide training data. The SMIRP framework is potentially applicable to all miRNA prediction systems and we expect substantial improvement in precision and specificity, while sustaining sensitivity, independent of the machine learning technique chosen.     

Systematic identification of microRNA functions by combining target prediction and expression profiling

Target predictions and validations are major obstacles facing microRNA (miRNA) researchers. Animal miRNA target prediction is challenging because of limited miRNA sequence complementarity to the targets. In addition, only a small number of predicted targets have been experimentally validated and the miRNA mechanism is poorly understood. Here we present a novel algorithm for animal miRNA target prediction. The algorithm combines relevant parameters for miRNA target recognition and heuristically assigns different weights to these parameters according to their relative importance. A score calculation scheme is introduced to reflect the strength of each parameter. We also performed microarray time course experiments to identify downregulated genes due to miRNA overexpression. The computational target prediction is combined with the miRNA transfection experiment to systematically identify the gene targets of human miR-124. miR-124 overexpression led to a significant downregulation of many cell cycle related genes. This may be the result of direct suppression of a few cell growth inhibitors at the early stage of miRNA overexpression, and these targeted genes were continuously suppressed over a long period of time. Our high-throughput approach can be generalized to globally identify the targets and functions of other miRNAs.     

Experimental strategies for microRNA target identification

MicroRNAs (miRNAs) are important regulators of eukaryotic gene expression in most biological processes. They act by guiding the RNAi-induced silencing complex (RISC) to partially complementary sequences in target mRNAs to suppress gene expression by a combination of translation inhibition and mRNA decay. The commonly accepted mechanism of miRNA targeting in animals involves an interaction between the 5'-end of the miRNA called the 'seed region' and the 3' untranslated region (3'-UTR) of the mRNA. Many target prediction algorithms are based around such a model, though increasing evidence demonstrates that targeting can also be mediated through sites other than the 3'-UTR and that seed region base pairing is not always required. The power and validity of such in silico data can be therefore hindered by the simplified rules used to represent targeting interactions. Experimentation is essential to identify genuine miRNA targets, however many experimental modalities exist and their limitations need to be understood. This review summarizes and critiques the existing experimental techniques for miRNA target identification.     

Weighted sequence motifs as an improved seeding step in microRNA target prediction algorithms

We present a new microRNA target prediction algorithm called TargetBoost, and show that the algorithm is stable and identifies more true targets than do existing algorithms. TargetBoost uses machine learning on a set of validated microRNA targets in lower organisms to create weighted sequence motifs that capture the binding characteristics between microRNAs and their targets. Existing algorithms require candidates to have (1) near-perfect complementarity between microRNAs' 5' end and their targets; (2) relatively high thermodynamic duplex stability; (3) multiple target sites in the target's 3' UTR; and (4) evolutionary conservation of the target between species. Most algorithms use one of the two first requirements in a seeding step, and use the three others as filters to improve the method's specificity. The initial seeding step determines an algorithm's sensitivity and also influences its specificity. As all algorithms may add filters to increase the specificity, we propose that methods should be compared before such filtering. We show that TargetBoost's weighted sequence motif approach is favorable to using both the duplex stability and the sequence complementarity steps. (TargetBoost is available as a Web tool from http://www.interagon.com/demo/.).