BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes

Tongxin Wang, Travis S. Johnson, Wei Shao, Zixiao Lu, Bryan R. Helm, Jie Zhang, Kun Huang

Research output: Contribution to journalArticle

Abstract

To fully utilize the power of single-cell RNA sequencing (scRNA-seq) technologies for identifying cell lineages and bona fide transcriptional signals, it is necessary to combine data from multiple experiments. We present BERMUDA (Batch Effect ReMoval Using Deep Autoencoders), a novel transfer-learning-based method for batch effect correction in scRNA-seq data. BERMUDA effectively combines different batches of scRNA-seq data with vastly different cell population compositions and amplifies biological signals by transferring information among batches. We demonstrate that BERMUDA outperforms existing methods for removing batch effects and distinguishing cell types in multiple simulated and real scRNA-seq datasets.

Original languageEnglish (US)
Article number165
JournalGenome Biology
Volume20
Issue number1
DOIs
StatePublished - Aug 12 2019

Fingerprint

RNA Sequence Analysis
RNA
learning
sequence analysis
cells
methodology
Cell Lineage
method
effect
removal
Transfer (Psychology)
Technology
Population
experiment

Keywords

  • Autoencoder
  • Batch effect
  • RNA-seq
  • Single cell
  • Transfer learning

ASJC Scopus subject areas

  • Ecology, Evolution, Behavior and Systematics
  • Genetics
  • Cell Biology

Cite this

BERMUDA : A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes. / Wang, Tongxin; Johnson, Travis S.; Shao, Wei; Lu, Zixiao; Helm, Bryan R.; Zhang, Jie; Huang, Kun.

In: Genome Biology, Vol. 20, No. 1, 165, 12.08.2019.

Research output: Contribution to journalArticle

Wang, Tongxin ; Johnson, Travis S. ; Shao, Wei ; Lu, Zixiao ; Helm, Bryan R. ; Zhang, Jie ; Huang, Kun. / BERMUDA : A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes. In: Genome Biology. 2019 ; Vol. 20, No. 1.
@article{2c714db25f2245afb847ab748028bb50,
title = "BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes",
abstract = "To fully utilize the power of single-cell RNA sequencing (scRNA-seq) technologies for identifying cell lineages and bona fide transcriptional signals, it is necessary to combine data from multiple experiments. We present BERMUDA (Batch Effect ReMoval Using Deep Autoencoders), a novel transfer-learning-based method for batch effect correction in scRNA-seq data. BERMUDA effectively combines different batches of scRNA-seq data with vastly different cell population compositions and amplifies biological signals by transferring information among batches. We demonstrate that BERMUDA outperforms existing methods for removing batch effects and distinguishing cell types in multiple simulated and real scRNA-seq datasets.",
keywords = "Autoencoder, Batch effect, RNA-seq, Single cell, Transfer learning",
author = "Tongxin Wang and Johnson, {Travis S.} and Wei Shao and Zixiao Lu and Helm, {Bryan R.} and Jie Zhang and Kun Huang",
year = "2019",
month = "8",
day = "12",
doi = "10.1186/s13059-019-1764-6",
language = "English (US)",
volume = "20",
journal = "Genome Biology",
issn = "1474-7596",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - BERMUDA

T2 - A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes

AU - Wang, Tongxin

AU - Johnson, Travis S.

AU - Shao, Wei

AU - Lu, Zixiao

AU - Helm, Bryan R.

AU - Zhang, Jie

AU - Huang, Kun

PY - 2019/8/12

Y1 - 2019/8/12

N2 - To fully utilize the power of single-cell RNA sequencing (scRNA-seq) technologies for identifying cell lineages and bona fide transcriptional signals, it is necessary to combine data from multiple experiments. We present BERMUDA (Batch Effect ReMoval Using Deep Autoencoders), a novel transfer-learning-based method for batch effect correction in scRNA-seq data. BERMUDA effectively combines different batches of scRNA-seq data with vastly different cell population compositions and amplifies biological signals by transferring information among batches. We demonstrate that BERMUDA outperforms existing methods for removing batch effects and distinguishing cell types in multiple simulated and real scRNA-seq datasets.

AB - To fully utilize the power of single-cell RNA sequencing (scRNA-seq) technologies for identifying cell lineages and bona fide transcriptional signals, it is necessary to combine data from multiple experiments. We present BERMUDA (Batch Effect ReMoval Using Deep Autoencoders), a novel transfer-learning-based method for batch effect correction in scRNA-seq data. BERMUDA effectively combines different batches of scRNA-seq data with vastly different cell population compositions and amplifies biological signals by transferring information among batches. We demonstrate that BERMUDA outperforms existing methods for removing batch effects and distinguishing cell types in multiple simulated and real scRNA-seq datasets.

KW - Autoencoder

KW - Batch effect

KW - RNA-seq

KW - Single cell

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85070662679&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070662679&partnerID=8YFLogxK

U2 - 10.1186/s13059-019-1764-6

DO - 10.1186/s13059-019-1764-6

M3 - Article

AN - SCOPUS:85070662679

VL - 20

JO - Genome Biology

JF - Genome Biology

SN - 1474-7596

IS - 1

M1 - 165

ER -