Can MiRBase Provide Positive Data for Machine Learning for the Detection of MiRNA Hairpins?

Saçar, Müşerref Duygu; Hamzeiy, Hamid; Allmer, Jens

doi:10.1515/jib-2013-215

Cited by 26 publications

(17 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previously published algorithms, based on two-class classification, have been evaluated using sensitivity, specificity, and accuracy as measures for their predictive power. We have previously shown that positive data derived from miRBase contains contaminating non-miRNAs [27]. Additionally, it is clear that the negative class cannot be established experimentally and that the proposed negative datasets are likely to contain miRNAs.…”

Section: Evaluation Methodsmentioning

confidence: 98%

“…Such approaches need positive (true miRNAs) and negative (non-miRNAs) data to become functional. Unfortunately, the positive data (usually derived from miRBase) may contain non-miRNAs [27]. Even worse, the negative class cannot be established experimentally and, therefore, most of these methods require the generation of an artificial negative class which may lead to problems [2].…”

Section: Computational Detection Of Pre-micrornasmentioning

confidence: 99%

See 1 more Smart Citation

Sequence Motif-Based One-Class Classifiers Can Achieve Comparable Accuracy to Two-Class Learners for Plant microRNA Detection

Yousef¹,

Allmer²,

Khalifa³

2015

JBiSE

Self Cite

View full text Add to dashboard Cite

microRNAs (miRNAs) are short nucleotide sequences expressed by a genome that are involved in post transcriptional modulation of gene expression. Since miRNAs need to be co-expressed with their target mRNA to observe an effect and since miRNAs and target interactions can be cooperative, it is currently not possible to develop a comprehensive experimental atlas of miRNAs and their targets. To overcome this limitation, machine learning has been applied to miRNA detection. In general binary learning (two-class) approaches are applied to miRNA discovery. These learners consider both positive (miRNA) and negative (non-miRNA) examples during the training process. One-class classifiers, on the other hand, use only the information for the target class (miRNA). The one-class approach in machine learning is gradually receiving more attention particularly for solving problems where the negative class is not well defined. This is especially true for miRNAs where the positive class can be experimentally confirmed relatively easy, but where it is not currently possible to call any part of a genome a non-miRNA. To do that, it should be co-expressed with all other possible transcripts of the genome, which currently is a futile endeavor. For machine learning, miRNAs need to be transformed into a feature vector and some currently used features like minimum free energy vary widely in the case of plant miRNAs. In this study it was our aim to analyze different methods applying one-class approaches and the effectiveness of motifbased features for prediction of plant miRNA genes. We show that the application of these oneclass classifiers is promising and useful for this kind of problem which relies only on sequencebased features such as k-mers and motifs comparing to the results from two-class classification. In * Corresponding author. M. Yousef et al. 685 some cases the results of one-class are, to our surprise, more accurate than results from two-class classifiers.

show abstract

Section: Evaluation Methodsmentioning

confidence: 98%

Section: Computational Detection Of Pre-micrornasmentioning

confidence: 99%

Sequence Motif-Based One-Class Classifiers Can Achieve Comparable Accuracy to Two-Class Learners for Plant microRNA Detection

Yousef¹,

Allmer²,

Khalifa³

2015

JBiSE

Self Cite

View full text Add to dashboard Cite

show abstract

“…At the end 7 different data sets with the most commonly used 10 [13] were produced for training. In the first dataset (1) 1600 positive and 800 negative examples were used.…”

Section: Data Miningmentioning

confidence: 99%

“…We recently showed that miRBase [11], the largest database for miRNAs, contains dubious examples for human and that, if all positive human miRNAs are used for classification, the accuracy is less than if the more stringently annotated examples from miRTarBase [12] are being used [13].…”

Section: Introductionmentioning

confidence: 99%

Data mining for microrna gene prediction: On the impact of class imbalance and feature number for microrna gene prediction

Saçar

Allmer

2013

2013 8th International Symposium on Health Informatics and Bioinformatics

Self Cite

View full text Add to dashboard Cite

Abstract-MicroRNAs (miRNAs) are small, non-coding RNAs which are involved in the posttranscriptional modulation of gene expression. Their short (18-24) single stranded mature sequences are involved in targeting specific genes. It turns out that experimental methods are limited and that it is difficult, if not impossible, to establish all miRNAs and their targets experimentally. Therefore, many tools for the prediction of miRNA genes and miRNA targets have been proposed. Most of these tools are based on machine learning methods and within that area mostly two-class classification is employed. Unfortunately, truly negative data is impossible to attain and only approximations of negative data are currently available. Also, we recently showed that the available positive data is not flawless. Here we investigate the impact of class imbalance on the learner accuracy and find that there is a difference of up to 50% between the best and worst precision and recall values. In addition, we looked at increasing number of features and found a curve maximizing at 0.97 recall and 0.91 precision with quickly decaying performance after inclusion of more than 100 features.

show abstract

“…Although other databases like miRTarBase [12], TarBase [13], and MirGeneDB [14] are available, positive data is generally derived from miRBase [15]. While negative data is of unknown quality, also positive data from miRBase contains questionable entries [14,16] and even MirGeneDB which filters miRBase entries is not free from questionable examples [17].…”

Section: Introductionmentioning

confidence: 99%

Categorization of species based on their microRNAs employing sequence motifs, information-theoretic sequence feature extraction, and k-mers

Yousef

Nigatu

Levy

et al. 2017

EURASIP J. Adv. Signal Process.

Self Cite

View full text Add to dashboard Cite

Background: Diseases like cancer can manifest themselves through changes in protein abundance, and microRNAs (miRNAs) play a key role in the modulation of protein quantity. MicroRNAs are used throughout all kingdoms and have been shown to be exploited by viruses to modulate their host environment. Since the experimental detection of miRNAs is difficult, computational methods have been developed. Many such tools employ machine learning for pre-miRNA detection, and many features for miRNA parameterization have been proposed. To train machine learning models, negative data is of importance yet hard to come by; therefore, we recently started to employ pre-miRNAs from one species as positive data versus another species' pre-miRNAs as negative examples based on sequence motifs and k-mers. Here, we introduce the additional usage of information-theoretic (IT) features. Results: Pre-miRNAs from one species were used as positive and another species' pre-miRNAs as negative training data for machine learning. The categorization capability of IT and k-mer features was investigated. Both feature sets and their combinations yielded a very high accuracy, which is as good as the previously suggested sequence motif and k-mer based method. However, for obtaining a high performance, a sufficiently large phylogenetic distance between the species and sufficiently high number of pre-miRNAs in the training set is required. To examine the contribution of the IT and k-mer features, an information gain-based feature ranking was performed. Although the top 3 are IT features, 80% of the top 100 features are k-mers. The comparison of all three individual approaches (motifs, IT, and k-mers) shows that the distinction of species based on their pre-miRNAs k-mers are sufficient. Conclusions: IT sequence feature extraction enables the distinction among species and is less computationally expensive than motif calculations. However, since IT features need larger amounts of data to have enough statistics for producing highly accurate results, future categorization into species can be effectively done using k-mers only. The biological reasoning for this is the existence of a codon bias between species which can, at least, be observed in exonic miRNAs. Future work in this direction will be the ab initio detection of pre-miRNA. In addition, prediction of pre-miRNA from RNA-seq can be done.

show abstract

Can MiRBase Provide Positive Data for Machine Learning for the Detection of MiRNA Hairpins?

Cited by 26 publications

References 27 publications

Sequence Motif-Based One-Class Classifiers Can Achieve Comparable Accuracy to Two-Class Learners for Plant microRNA Detection

Sequence Motif-Based One-Class Classifiers Can Achieve Comparable Accuracy to Two-Class Learners for Plant microRNA Detection

Data mining for microrna gene prediction: On the impact of class imbalance and feature number for microrna gene prediction

Categorization of species based on their microRNAs employing sequence motifs, information-theoretic sequence feature extraction, and k-mers

Contact Info

Product

Resources

About