MiRTDL: A Deep Learning Approach for miRNA Target Prediction

Cheng, Shuang; Guo, Maozu; Wang, Chuyu; Liu, Xiaoyan; Liu, Yang; Wu, Xuejian

doi:10.1109/tcbb.2015.2510002

Cited by 71 publications

(39 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although performance may increase significantly by changing other parameters, for example, the batch size and the number of units in the remaining layers, our results support the importance of spatial features in miRNA target prediction. One previous study, the miRTDL study, applied CNN in miRNA target prediction, however, the authors only used CNN as a classifier (Cheng, et al, 2016). The miTAR model is the first to use CNN to capture the potential spatial features directly from the sequences of miRNAs and genes.…”

Section: Discussionmentioning

confidence: 99%

“…Beside DeepMirTar and miRAW, we compared our approach to two other published deep learning methods: deepTarget and miRTDL (Cheng, et al, 2016;Lee, et al, 2016). They were developed earlier, hence, the training datasets they used are much smaller.…”

Section: Performance Comparison With Earlier Studies Using Test Datasetsmentioning

confidence: 99%

“…Recently, several deep learning methods were developed to handle unknown features and to improve the accuracy of prediction. MiRTDL (Cheng, et al, 2016) uses a convolutional neural network (CNN) to capture training sets features. Although CNNs can automatically assess feature importance, miRTDL is still based on known features.…”

mentioning

confidence: 99%

See 2 more Smart Citations

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

Zhao

Barbazuk

et al. 2020

Preprint

View full text Add to dashboard Cite

microRNAs (miRNAs) are a major type of small RNA that alter gene expression at the posttranscriptional or translational level. They have been shown to play important roles in a wide range of biological processes. Many computational methods have been developed to predict targets of miRNAs in order to understand miRNAs' function. However, the majority of the methods depend on a set of predefined features that require considerable effort and resources to compute, and these methods often do not effectively on the prediction of miRNA targets. Therefore, we developed a novel hybrid deep learningbased approach that is capable to predict miRNA targets at a higher accuracy. Our approach integrates two deep learning methods: convolutional neural networks (CNNs) that excel in learning spatial features, and recurrent neural networks (RNNs) that discern sequential features. By combining CNNs and RNNs, our approach has the advantages of learning both the intrinsic spatial and sequential features of miRNA:target. The inputs for the approach are raw sequences of miRNA and gene sequences. Data from two latest miRNA target prediction studies were used in our study: the DeepMirTar dataset and the miRAW dataset. Two models were obtained by training on the two datasets separately. The models achieved a higher accuracy than the methods developed in the previous studies: 0.9787 vs. 0.9348 for the DeepMirTar dataset; 0.9649 vs. 0.935 for the miRAW dataset. We also calculated a series of model evaluation metrics including sensitivity, specificity, F-score and Brier Score. Our approach consistently outperformed the current methods. In addition, we compared our approach with earlier developed deep learning methods, resulting in an overall better performance. Lastly, a unified model for both datasets was developed with an accuracy of 0.9545. We named the unified model miTAR for miRNA target prediction. The source code and executable are available at https://github.com/tjgu/miTAR. Advances in our understanding of the interactions between miRNAs and their targets has led to the development of many computational methods/tools to predict miRNA targets. The majority of these tools are based on common features of the miRNA:target interaction. Four features are widely used: sequence complement (especially in the seed region that is generally defined as a 6 or 7 nts sequence starting at the second or third nucleotide(nt) of the miRNA sequence), thermodynamic stability, target site accessibility and sequence conservation among species (Peterson, et al., 2014). Several widely used tools have been developed based on these features. For example, miRanda (Enright, et al., 2003) relies on sequence complementarity and binding energy; TargetScanS (Lewis, et al., 2005) relies on sequence complementarity in seed region; while PITA (Kertesz, et al., 2007) relies on target site accessibility. However, miRNA targets predicted by different methods and tools are inconsistent with one another. Furthermore, using known features limits the ability to predict novel o...

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Performance Comparison With Earlier Studies Using Test Datasetsmentioning

confidence: 99%

See 1 more Smart Citation

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

Zhao

Barbazuk

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The majority of prediction tools are based on the assumption that it is the 12 miRNA seed region -generally defined as a 6 to 8 nucleotide sequence starting at the 13 first or second nucleotide -that contains almost all the important interactions between 14 a miRNA and its target and their focus is on these canonical sites. This seed-centric 15 view has been supported by structural studies [5] and a widely cited report [6] that 16 investigated the importance of other (non-canonical) regions within a miRNA and 17 concluded their contributions had relatively low relevance compared to the (canonical) 18 seed region. However, more recent studies have revealed that many relevant targets are 19 implemented via non-canonical binding and involve nucleotides outside the seed region, 20 indicating that the entire miRNA should be considered in target prediction 21 algorithms [3,7,8].…”

mentioning

confidence: 85%

miRAW: A deep learning approach to predict miRNA targets by analyzing whole miRNA transcripts

Plà

Zhong

Rayner

2017

Preprint

View full text Add to dashboard Cite

MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression by binding to partially complementary regions within the 3'UTR of their target genes. Computational methods play an important role in target prediction and assume that the miRNA "seed region" (nt 2 to 8) is required for functional targeting, but typically only identify 80% of known bindings. Recent studies have highlighted a role for the entire miRNA, suggesting that a more flexible methodology is needed.We present a novel approach for miRNA target prediction based on Deep Learning (DL) which, rather than incorporating any knowledge (such as seed regions), investigates the entire miRNA and 3'UTR mRNA nucleotides to learn a uninhibited set of feature descriptors related to the targeting process.We collected more than 150,000 experimentally validated homo sapiens miRNA:gene targets and cross referenced them with different CLIP-Seq, CLASH and iPAR-CLIP datasets to obtain 20,000 validated miRNA:gene exact target sites. Using this data, we implemented and trained a deep neural network -composed of autoencoders and a feed-forward network -able to automatically learn features describing miRNA-mRNA interactions and assess functionality. Predictions were then refined using information such as site location or site accessibility energy.In a comparison using independent datasets, our DL approach consistently outperformed existing prediction methods, recognizing the seed region as a common feature in the targeting process, but also identifying the role of pairings outside this region. Thermodynamic analysis also suggests that site accessibility plays a role in targeting but that it cannot be used as a sole indicator for functionality.Data and source code available at: https://bitbucket.org/account/user/bipous/projects/MIRAW Author summary microRNAs are small RNA molecules that regulate biological processes by binding to the 3'UTR of a gene and their dysregulation is associated with several diseases. Computationally predicting these targets remains a challenge as they only partially match their target and so there can be hundreds of targets for a single microRNA.Current tools assume that most of the knowledge defining a microRNA-gene interaction can be captured by analysing the binding produced in the seed region (≈ the first 8nt in the miRNA). However, recent studies show that the whole microRNA can be important and form non-canonical targets. Here, we use a target prediction methodology miRAW: deep learning for miRNA target prediction -Pla et al.

show abstract

“…They used the model to detect a unified representation of latent features, capture both intra- and cross- modality correlations, and to identify key genes that may play distinct roles in the pathogenesis between different cancer subtypes. Cheng et al [16] designed a miRNA prediction algorithm based on convolutional neural network (CNN). The CNN automatically extracts essential information from the input data while the exact miRNA target mechanisms are not well known.…”

Section: Introductionmentioning

confidence: 99%

Identify Huntington’s disease associated genes based on restricted Boltzmann machine with RNA-seq data

et al. 2017

View full text Add to dashboard Cite

BackgroundPredicting disease-associated genes is helpful for understanding the molecular mechanisms during the disease progression. Since the pathological mechanisms of neurodegenerative diseases are very complex, traditional statistic-based methods are not suitable for identifying key genes related to the disease development. Recent studies have shown that the computational models with deep structure can learn automatically the features of biological data, which is useful for exploring the characteristics of gene expression during the disease progression.ResultsIn this paper, we propose a deep learning approach based on the restricted Boltzmann machine to analyze the RNA-seq data of Huntington’s disease, namely stacked restricted Boltzmann machine (SRBM). According to the SRBM, we also design a novel framework to screen the key genes during the Huntington’s disease development. In this work, we assume that the effects of regulatory factors can be captured by the hierarchical structure and narrow hidden layers of the SRBM. First, we select disease-associated factors with different time period datasets according to the differentially activated neurons in hidden layers. Then, we select disease-associated genes according to the changes of the gene energy in SRBM at different time periods.ConclusionsThe experimental results demonstrate that SRBM can detect the important information for differential analysis of time series gene expression datasets. The identification accuracy of the disease-associated genes is improved to some extent using the novel framework. Moreover, the prediction precision of disease-associated genes for top ranking genes using SRBM is effectively improved compared with that of the state of the art methods.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-017-1859-6) contains supplementary material, which is available to authorized users.

show abstract

MiRTDL: A Deep Learning Approach for miRNA Target Prediction

Cited by 71 publications

References 32 publications

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

miRAW: A deep learning approach to predict miRNA targets by analyzing whole miRNA transcripts

Identify Huntington’s disease associated genes based on restricted Boltzmann machine with RNA-seq data

Contact Info

Product

Resources

About