2019
DOI: 10.1101/837302
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

scGAIN: Single Cell RNA-seq Data Imputation using Generative Adversarial Networks

Abstract: Single cell RNA sequencing (scRNA-seq) provides a rich view into the heterogeneity underlying a cell population. However single-cell data are usually noisy and very sparse due to the presence of dropout genes. In this work we propose an approach to impute missing gene expressions in single cell data using generative adversarial networks (GANs). By learning an approximate distribution of the data, our approach, scGAIN, can impute dropouts in simulated and real single cell data. The work in this paper discusses … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
7
0

Year Published

2020
2020
2023
2023

Publication Types

Select...
6
2

Relationship

0
8

Authors

Journals

citations
Cited by 10 publications
(7 citation statements)
references
References 21 publications
0
7
0
Order By: Relevance
“…The raw scRNA-seq has a much higher zero expression rate than bulk RNA-seq (49.1% versus 14.8%) and shares fewest DEGs with bulk samples (Figure 4A ). After imputation, the number of DEGs is increased toward the DEGs numbers of bulk samples (except the four other deep learning-based methods, AutoImpute ( 48 ), DCA ( 6 ), DeepImpute ( 8 ) and scGAIN ( 49 ), which detect much fewer DEGs than raw data). Especially highlighted by the yellow and green bars in Figure 4A , the AutoImpute and scGAIN detect significantly fewer DEGs and show the poorest agreement with other methods.…”
Section: Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…The raw scRNA-seq has a much higher zero expression rate than bulk RNA-seq (49.1% versus 14.8%) and shares fewest DEGs with bulk samples (Figure 4A ). After imputation, the number of DEGs is increased toward the DEGs numbers of bulk samples (except the four other deep learning-based methods, AutoImpute ( 48 ), DCA ( 6 ), DeepImpute ( 8 ) and scGAIN ( 49 ), which detect much fewer DEGs than raw data). Especially highlighted by the yellow and green bars in Figure 4A , the AutoImpute and scGAIN detect significantly fewer DEGs and show the poorest agreement with other methods.…”
Section: Resultsmentioning
confidence: 99%
“…During the submission of this work, another work using GANs for scRNA-seq imputation (referred to as scGAIN) was posted in BioRxiv ( 49 ). We added it in all evaluations and downstream analyses in our work and found that scGAIN is not as good as scIGANs in all evaluations.…”
Section: Discussionmentioning
confidence: 99%
“…Low reverse transcription efficiency in single-cell RNA-seq data. [13][14][15][16][17][18][19][20][21] Proteomics data…”
Section: Genomics Datamentioning
confidence: 99%
“…Many methods and tools are currently available for solving the dropout issues of scRNA-seq data (Pierson and Yau, 2015;Azizi et al, 2017;Lin et al, 2017;Chen and Zhou, 2018;Gong et al, 2018;Huang et al, 2018;Li and Li, 2018;Ronen and Akalin, 2018;van Dijk et al, 2018;Amodio et al, 2019;Arisdakessian et al, 2019;Eraslan et al, 2019;Gunady et al, 2019;Peng et al, 2019;Tracy et al, 2019;Wagner et al, 2019;Badsha et al, 2020;Marouf et al, 2020;Xu et al, 2020). MAGIC recovers dropout events by using diffusion geometry to share similarities across cells (van Dijk et al, 2018).…”
Section: Recover Dropout Events In Single-cell Transcriptome Profilesmentioning
confidence: 99%