Application of Deep Learning on Single-Cell RNA Sequencing Data Analysis: A Review

Abstract

Single-cell RNA sequencing (scRNA-seq) has become a routinely used technique to quantify the gene expression profile of thousands of single cells simultaneously. Analysis of scRNA-seq data plays an important role in the study of cell states and phenotypes, and has helped elucidate biological processes, such as those occurring during the development of complex organisms, and improved our understanding of disease states, such as cancer, diabetes, and coronavirus disease 2019 (COVID-19). Deep learning, a recent advance of artificial intelligence that has been used to address many problems involving large datasets, has also emerged as a promising tool for scRNA-seq data analysis, as it has a capacity to extract informative and compact features from noisy, heterogeneous, and high-dimensional scRNA-seq data to improve downstream analysis. The present review aims at surveying recently developed deep learning techniques in scRNA-seq data analysis, identifying key steps within the scRNA-seq data analysis pipeline that have been advanced by deep learning, and explaining the benefits of deep learning over more conventional analytic tools. Finally, we summarize the challenges in current deep learning approaches faced within scRNA-seq data and discuss potential directions for improvements in deep learning algorithms for scRNA-seq data analysis.

Data normalization for addressing the challenges in the analysis of single-cell transcriptomic datasets

“…Deep learning, a subclass of machine learning, has been recently used to analyze high-throughput omics data, including scRNA-seq [ 119 ]. Deep learning consists of neural network architectures to discover latent and informative patterns in complex data incorporating thousands of trainable parameters and finds transformations that can effectively normalize counts preserving biological information [ 120 ].…”

From cancer big data to treatment: Artificial intelligence in cancer research

“…Deep learning has emerged as a powerful approach in cancer research, particularly in the analysis and prediction of gene expression data obtained from RNAseq and scRNAseq technologies. [57][58][59] Some of the most commonly used neural network architectures in the field include fully connected neural networks, convolutional neural networks (CNNs), recurrent neural networks (RNNs) and graph neural networks (GNNs). [60][61][62] RNNs are a type of neural network that includes recurrent connections between the neuron units, providing the network with a memory capability.…”

“…Deep learning is a subset of ML that involves artificial neural networks with multiple layers to automatically extract complex patterns from data. Deep learning has emerged as a powerful approach in cancer research, particularly in the analysis and prediction of gene expression data obtained from RNAseq and scRNAseq technologies 57–59 . Some of the most commonly used neural network architectures in the field include fully connected neural networks, convolutional neural networks (CNNs), recurrent neural networks (RNNs) and graph neural networks (GNNs) 60–62 .…”

Inferring Disease Progressive Stages in Single-Cell Transcriptomics Using Weakly-Supervised Deep Learning Approach

“…The deep learning has been widely implemented in various single-cell data analyses, including data imputation [11], doublet identification [12], dimensional reduction [13], batch effect corrections [14], and cell type annotations [15, 16]. However, there is a limited application of the deep learning to the inference of disease progression of individual cells [17]. One of major challenges may be difficulty to train the model and regress continuous disease progression levels from the binary diagnosis information (e.g.…”