Exploring the optimization of autoencoder design for imputing single-cell RNA sequencing data
Open Access
- 1 January 2023
- journal article
- research article
- Published by American Association for the Advancement of Science (AAAS) in Computational and Structural Biotechnology Journal
- Vol. 21, 4079-4095
Abstract
Autoencoders are the backbones of many imputation methods that aim to relieve the sparsity issue in single-cell RNA sequencing (scRNA-seq) data. The imputation performance of an autoencoder relies on both the neural network architecture and the hyperparameter choice. So far, literature in the single-cell field lacks a formal discussion on how to design the neural network and choose the hyperparameters. Here, we conducted an empirical study to answer this question. Our study used many real and simulated scRNA-seq datasets to examine the impacts of the neural network architecture, the activation function, and the regularization strategy on imputation accuracy and downstream analyses. Our results show that (i) deeper and narrower autoencoders generally lead to better imputation performance; (ii) the sigmoid and tanh activation functions consistently outperform other commonly used functions including ReLU; (iii) regularization improves the accuracy of imputation and downstream cell clustering and DE gene analyses. Notably, our results differ from common practices in the computer vision field regarding the activation function and the regularization strategy. Overall, our study offers practical guidance on how to optimize the autoencoder design for scRNA-seq data imputation.Keywords
All Related Versions
Funding Information
- National Institutes of Health
- National Sleep Foundation
This publication has 41 references indexed in Scilit:
- DrImpute: imputing dropout events in single cell RNA sequencing dataBMC Bioinformatics, 2018
- An accurate and robust imputation method scImpute for single-cell RNA-seq dataNature Communications, 2018
- Missing data and technical variability in single-cell RNA-sequencing experimentsBiostatistics, 2017
- Using neural networks for reducing the dimensions of single-cell RNA-Seq dataNucleic Acids Research, 2017
- Bayesian Inference for Single-cell Clustering and ImputingGenomics and Computational Biology, 2017
- MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing dataGenome Biology, 2015
- ZIFA: Dimensionality reduction for zero-inflated single-cell gene expression analysisGenome Biology, 2015
- Deep learningNature, 2015
- mRNA-Seq whole-transcriptome analysis of a single cellNature Methods, 2009
- Backpropagation through time: what it does and how to do itProceedings of the IEEE, 1990