Dataset Information

HuntMi: an efficient and taxon-specific approach in pre-miRNA identification.

ABSTRACT:

Background

Machine learning techniques are known to be a powerful way of distinguishing microRNA hairpins from pseudo hairpins and have been applied in a number of recognised miRNA search tools. However, many current methods based on machine learning suffer from some drawbacks, including not addressing the class imbalance problem properly. It may lead to overlearning the majority class and/or incorrect assessment of classification performance. Moreover, those tools are effective for a narrow range of species, usually the model ones. This study aims at improving performance of miRNA classification procedure, extending its usability and reducing computational time.

Results

We present HuntMi, a stand-alone machine learning miRNA classification tool. We developed a novel method of dealing with the class imbalance problem called ROC-select, which is based on thresholding score function produced by traditional classifiers. We also introduced new features to the data representation. Several classification algorithms in combination with ROC-select were tested and random forest was selected for the best balance between sensitivity and specificity. Reliable assessment of classification performance is guaranteed by using large, strongly imbalanced, and taxon-specific datasets in 10-fold cross-validation procedure. As a result, HuntMi achieves a considerably better performance than any other miRNA classification tool and can be applied in miRNA search experiments in a wide range of species.

Conclusions

Our results indicate that HuntMi represents an effective and flexible tool for identification of new microRNAs in animals, plants and viruses. ROC-select strategy proves to be superior to other methods of dealing with class imbalance problem and can possibly be used in other machine learning classification tasks. The HuntMi software as well as datasets used in the research are freely available at http://lemur.amu.edu.pl/share/HuntMi/.

SUBMITTER: Gudys A

PROVIDER: S-EPMC3686668 | biostudies-literature | 2013 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

HuntMi: an efficient and taxon-specific approach in pre-miRNA identification.

Gudyś Adam A Szcześniak Michał Wojciech MW Sikora Marek M Makałowska Izabela I

BMC bioinformatics 20130305

<h4>Background</h4>Machine learning techniques are known to be a powerful way of distinguishing microRNA hairpins from pseudo hairpins and have been applied in a number of recognised miRNA search tools. However, many current methods based on machine learning suffer from some drawbacks, including not addressing the class imbalance problem properly. It may lead to overlearning the majority class and/or incorrect assessment of classification performance. Moreover, those tools are effective for a na ...[more]

PMID: 23497112

Similar Datasets

Project description:BackgroundMicroRNAs (miRNAs) are a set of short (19∼24 nt) non-coding RNAs that play significant roles as posttranscriptional regulators in animals and plants. The ab initio prediction methods show excellent performance for discovering new pre-miRNAs. While most of these methods can distinguish real pre-miRNAs from pseudo pre-miRNAs, few can predict the positions of miRNAs. Among the existing methods that can also predict the miRNA positions, most of them are designed for mammalian miRNAs, including human and mouse. Minority of methods can predict the positions of plant miRNAs. Accurate prediction of the miRNA positions remains a challenge, especially for plant miRNAs. This motivates us to develop MaturePred, a machine learning method based on support vector machine, to predict the positions of plant miRNAs for the new plant pre-miRNA candidates.Methodology/principal findingsA miRNA:miRNA* duplex is regarded as a whole to capture the binding characteristics of miRNAs. We extract the position-specific features, the energy related features, the structure related features, and stability related features from real/pseudo miRNA:miRNA* duplexes. A set of informative features are selected to improve the prediction accuracy. Two-stage sample selection algorithm is proposed to combat the serious imbalance problem between real and pseudo miRNA:miRNA* duplexes. The prediction method, MaturePred, can accurately predict plant miRNAs and achieve higher prediction accuracy compared with the existing methods. Further, we trained a prediction model with animal data to predict animal miRNAs. The model also achieves higher prediction performance. It further confirms the efficiency of our miRNA prediction method.ConclusionsThe superior performance of the proposed prediction model can be attributed to the extracted features of plant miRNAs and miRNA*s, the selected training dataset, and the carefully selected features. The web service of MaturePred, the training datasets, the testing datasets, and the selected features are freely available at http://nclab.hit.edu.cn/maturepred/.

Project description:Many environmental and biomedical biomonitoring and detection studies aim to explore the presence of specific organisms or gene functionalities in microbiome samples. In such cases, when the study hypotheses can be answered with the exploration of a small number of genes, a targeted PCR-approach is appropriate. However, due to the complexity of environmental microbial communities, the design of specific primers is challenging and can lead to non-specific results. We designed PhyloPrimer, the first user-friendly platform to semi-automate the design of taxon-specific oligos (i.e., PCR primers) for a gene of interest. The main strength of PhyloPrimer is the ability to retrieve and align GenBank gene sequences matching the user's input, and to explore their relationships through an online dynamic tree. PhyloPrimer then designs oligos specific to the gene sequences selected from the tree and uses the tree non-selected sequences to look for and maximize oligo differences between targeted and non-targeted sequences, therefore increasing oligo taxon-specificity (positive/negative consensus approach). Designed oligos are then checked for the presence of secondary structure with the nearest-neighbor (NN) calculation and the presence of off-target matches with in silico PCR tests, also processing oligos with degenerate bases. Whilst the main function of PhyloPrimer is the design of taxon-specific oligos (down to the species level), the software can also be used for designing oligos to target a gene without any taxonomic specificity, for designing oligos from preselected sequences and for checking predesigned oligos. We validated the pipeline on four commercially available microbial mock communities using PhyloPrimer to design genus- and species-specific primers for the detection of Streptococcus species in the mock communities. The software performed well on these mock microbial communities and can be found at https://www.cerealsdb.uk.net/cerealgenomics/phyloprimer.

Dataset Information

HuntMi: an efficient and taxon-specific approach in pre-miRNA identification.

Background

Results

Conclusions

Publications

HuntMi: an efficient and taxon-specific approach in pre-miRNA identification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets