Deep Learning for Genomics: A Concise Overview

Abstract

This data explosion driven by advancements in genomic research, such as high-throughput sequencing techniques, is constantly challenging conventional methods used in genomics. In parallel with the urgent demand for robust algorithms, deep learning has succeeded in a variety of fields such as vision, speech, and text processing. Yet genomics entails unique challenges to deep learning since we are expecting from deep learning a superhuman intelligence that explores beyond our knowledge to interpret the genome. A powerful deep learning model should rely on insightful utilization of task-specific knowledge. In this paper, we briefly discuss the strengths of different deep learning models from a genomic perspective so as to fit each particular task with a proper deep architecture, and remark on practical considerations of developing modern deep learning architectures for genomics. We also provide a concise review of deep learning applications in various aspects of genomic research, as well as pointing out current challenges and potential research directions for future genomics applications.

Removing Confounding Factors Associated Weights in Deep Neural Networks Improves the Prediction Accuracy for Healthcare Applications

“…Readers can refer to comprehensive reviews on how the deep learning can be applied to healthcare and biomedical areas. 8,[22][23][24] In this section, we will mainly discuss the related work of our paper in the methodological perspective.…”

Identifying viruses from metagenomic data using deep learning

“…Recently, a few studies developed deep learning methods to solve different problems in computational biology, such as predicting the sequence specificity of protein binding using deep learning [1,48,32,43], predicting the effect of noncoding variants based on different model structures [49,31], predicting chromatin accessibility [9], calling genetic variants from millions of short reads in NGS samples [30], predicting methylation quantitative trait loci [47], identifying evolutionarily conserved sequences [17], and identification of enhancer and promoter [18] and their interactions [39]. See [46] for detailed reviews of deep learning applications in genomic research. Those methods have shown remarkable improvements over the traditional machine learning-based and statistical inferences-based models.…”

Analysis of three dimensional potential problems in non-homogeneous media with physics-informed deep collocation method using material transfer learning and sensitivity analysis

“…Various variants such as multilayer perceptron (MLP), convolutional neural networks (CNN) and recurrent/recursive neural networks (RNN) [9] have been developed and applied to e.g. image processing [10,11], object detection [12,13], speech recognition [14,15], biology [16,17] and even finance [18,19]. Deep learning can learn features from data automatically, and the features can be used to get the approximation of solutions to differential equations [20].…”