DP-miRNA: An improved prediction of precursor microRNA using deep learning model

Thomas, Jaya; Thomas, Sonia; Sael, Lee

doi:10.1109/bigcomp.2017.7881722

Cited by 18 publications

(17 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These include microPred (Batuwita and Palade, 2009b), triplet-SVM (Xue et al, 2005), and miRBoost (Tempel et al, 2015), which use different numbers of human and cross-species manual features to identify miRNAs as inputs to a support vector machine(SVM); and MiPred (Jiang et al, 2007), which selects a set of mixed features, including the minimum free energy (MFE), the local contiguous triplet structure composition, dinucleotide shuffling, and the P-values of randomization tests, to construct a random forest classifier to identify miRNAs. The context-sensitive hidden Markov model (CSHMM) method (Agarwal et al, 2010) predicts miRNAs by filtering the human dataset; whereas M0iRANN (Rahman et al, 2012), DP-miRNA (Thomas et al, 2017), and BP (Jiang et al, 2016) extracted 98 features as inputs to their neural networks. These methods use hand-crafted features as inputs to the model, including pre-miRNA structural and folding energy information such as dinucleotide and trinucleotide pair frequency, loop and sequence length, MFE, and melting temperature.…”

Section: Introductionmentioning

confidence: 99%

CL-PMI: A Precursor MicroRNA Identification Method Based on Convolutional and Long Short-Term Memory Networks

Wang

Dong

et al. 2019

Front. Genet.

View full text Add to dashboard Cite

MicroRNAs (miRNAs) are the major class of gene-regulating molecules that bind mRNAs. They function mainly as translational repressors in mammals. Therefore, how to identify miRNAs is one of the most important problems in medical treatment. Many known pre-miRNAs have a hairpin ring structure containing more structural features, and it is difficult to identify mature miRNAs because of their short length. Therefore, most research focuses on the identification of pre-miRNAs. Most computational models rely on manual feature extraction to identify pre-miRNAs and do not consider the sequential and spatial characteristics of pre-miRNAs, resulting in a loss of information. As the number of unidentified pre-miRNAs is far greater than that of known pre-miRNAs, there is a dataset imbalance problem, which leads to a degradation of the performance of pre-miRNA identification methods. In order to overcome the limitations of existing methods, we propose a pre-miRNA identification algorithm based on a cascaded CNN-LSTM framework, called CL-PMI. We used a convolutional neural network to automatically extract features and obtain pre-miRNA spatial information. We also employed long short-term memory (LSTM) to capture time characteristics of pre-miRNAs and improve attention mechanisms for long-term dependence modeling. Focal loss was used to improve the dataset imbalance. Compared with existing methods, CL-PMI achieved better performance on all datasets. The results demonstrate that this method can effectively identify pre-miRNAs by simultaneously considering their spatial and sequential information, as well as dealing with imbalance in the datasets.

show abstract

Section: Introductionmentioning

confidence: 99%

CL-PMI: A Precursor MicroRNA Identification Method Based on Convolutional and Long Short-Term Memory Networks

Wang

Dong

et al. 2019

Front. Genet.

View full text Add to dashboard Cite

show abstract

“…Previous work on miRNA and pre-miRNA identification has been based on handcrafted rules (MIReNA [31]) or machine learning. Machine learning methods have been increasingly popular during the last decade and demonstrated to be the most promising, with tools such as HuntMi [15], miRBoost [42], CSHMM [1], microPred [4], miPred [34], triplet-SVM [46], Mirann [37], DP-miRNA [43], deepMiR-Gene [35]. They predict secondary structure, which forms intricate base-pairing interactions within RNA sequence (see Fig.…”

Section: Introductionmentioning

confidence: 99%

“…1AB for an example), with standard methods such as RNAfold [18], GTfold [32], and CyloFold [6]. Then they extract a lot of handcrafted features, some of which are Watson-Crick nucleotide pairing (A-U, C-G), loop length [43,42], sequence length [43], dinucleotide pair frequencies [43,20,42,4,34], trinucleotide pair frequencies (constituting 64 features) [43,20], melting temperature [43], mini-mum free energy [44,43,9]. These features are used as inputs to machine learning methods such as support vector machines (SVM) [42,4,46], random forests [34], neural networks [37,43,35,20] and hidden Markov models [1].…”

Section: Introductionmentioning

confidence: 99%

“…[20] uses 98 of the aforementioned features for the input of their neural network. DP-miRNA [43] also uses neural network with 58 extracted characteristic features based on sequence composition and secondary structure predicted by RNAfold software, folding measures including various formulations of physical energy. One of the most feature-rich methods is Ref.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Precursor microRNA Identification Using Deep Convolutional Neural Networks

Golkov

et al. 2018

Preprint

View full text Add to dashboard Cite

Precursor microRNA (pre-miRNA) identification is the basis for identifying microRNAs (miRNAs), which have important roles in post-transcriptional regulation of gene expression. In this paper, we propose a deep learning method to identify whether a small non-coding RNA sequence is a pre-miRNA or not. We outperform state-ofthe-art methods on three benchmark datasets, namely the human, cross-species, and new datasets. The key of our method is to use a matrix representation of predicted secondary structure as input to a 2D convolutional network. The neural network extracts optimized features automatically instead of using a large number of handcrafted features as most existing methods do. Code and results are available at https://github.com/peace195/ miRNA-identification-conv2D. * Work done as an exchange student of ERASMUS+ Key Action 1 program between HUST and TUM.

show abstract

“…In [7] a deep belief neural network (deepBN) for identifying pre-miRNA sequences was proposed. This model has an unsupervised stage with hidden layers pre-trained as restricted Boltzmann machines, followed by a supervised tuning of the network.…”

mentioning

confidence: 99%