Deep learning of the tissue-regulated splicing code

By Michael K. K. Leung, Hui Yuan Xiong, Leo J. Lee, Brendan J. Frey. 2014

455 citations recorded by the source corpus.

Motivation: Alternative splicing (AS) is a regulated process that directs the generation of different transcripts from single genes. A computational model that can accurately predict splicing patterns based on genomic features and cellular context is highly desirable, both in understanding this widespread phenomenon, and in exploring the effects of genetic variations on AS. Methods: Using a deep neural network, we developed a model inferred from mouse RNA-Seq data that can predict splicing patterns in individual tissues and differences in splicing patterns across tissues. Our architecture uses hidden variables that jointly represent features in genomic sequences and tissue types when making predictions. A graphics processing unit was used to greatly reduce the training time of our models with millions of parameters. Results: We show that the deep architecture surpasses the performance of the previous Bayesian method for predicting AS patterns. With the proper optimization procedure and selection of hyperparameters, we demonstrate that deep architectures can be beneficial, even with a moderately sparse dataset. An analysis of what the model has learned in terms of the genomic features is presented. Contact: frey@psi.toronto.edu Supplementary information: Supplementary data are available at Bioinformatics online.

Subfields: Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

A deep learning framework for financial time series using stacked autoencoders and long-short term memory (2017, neutral): “For example, deep learning-based methods have dramatically improved the state-of-the-art in image recognition [12, 39–41], speech recognition[42–44], language translation[45, 46] and many other areas such as drug discovery [47] and genomics [48, 49].”
Understanding deep convolutional networks (2016, background): “They give state of the art performances for image classification with thousands of complex classes [19], speech recognition [17], bio-medical applications [22], natural language understanding [30], and in many other domains.”
Deep Patient: An Unsupervised Representation to Predict the Future of Patients from the Electronic Health Records (2016, neutral): “Deep learning was recently applied to medicine and genomics to reconstruct brain circuits 35 and to predict the activity of potential drug molecules 36 , the effects of mutations in non-coding DNA on gene expressions 37,38 , and the sequence specificities of DNA and RNA-binding proteins 39 .”
Deep learning for computational biology (2016, positive): “…relatively low number of 10,700 training samples in combination with the model complexity, this method achieved substantially higher prediction accuracy of splicing activity compared to simpler approaches, and in particular was able to identify rare mutations implicated in splicing misregulation.”
Computer vision for high content screening (2016, background): “Examples of biological applications of deep learning include predicting the effects of genetic variants on RNA splicing (Leung et al ., 2014; Xiong et al ., 2015) and drug profiling and discovery (Ma et al ., 2015; Ramsundar et al ., 2015).”
Deep learning in bioinformatics: introduction, application, and perspective in big data era (2019, methodology): “For recurrent neural networks, the more advanced architectures help people deal with the gradient vanishing or explosion issue and accelerate the execution of RNN.”
Gene expression inference with deep learning (2015, background): “With these advances, deep learning has achieved state-of-the-art performances on a wide range of applications, both in traditional machine learning tasks such as computer vision [13], natural language processing [14], speech recognition [15], and in natural science applications such as exotic partic”
Using deep Q-learning to understand the tax evasion behavior of risk-averse firms (2018, background): “, 2011), biology (Leung et al., 2014), analysis of financial trading (Krauss et al.”
Deep Learning for Intelligent Wireless Networks: A Comprehensive Survey (2018, background): “many real-world control scenarios, such as voice recognition [2], [3], image recognition [4], [5], [6], [7], semantic analysis [8], [9], language interpretation [10], [11], game control [12], drug discovery [13], and biomedical sciences [14], [15], [16], etc.”
Integrative deep models for alternative splicing (2017, methodology): “DNN-LMH 203 was designed according to the architecture and hyperparameters described in [9].”
Development and evaluation of a deep learning model for protein–ligand binding affinity prediction (2017, methodology): “Deep learning has been relatively widely used by the bioinformatics (Leung et al. , 2014; Alipanahi et al. , 2015; Park and Kellis, 2015; Jurtz et al. , 2017; Jim´enez et al. , 2017) and computational biology community (Angermueller et al. , 2016).”