Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics
Abstract Background Single-cell RNA sequencing is a powerful tool for characterizing cellular heterogeneity in gene expression. However, high variability and a large number of zero counts present challenges for analysis and interpretation. There is substantial controversy over the origins and proper...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2020-07-01
|
Series: | Genome Biology |
Subjects: | |
Online Access: | http://link.springer.com/article/10.1186/s13059-020-02103-2 |
id |
doaj-9044e5bb4d9a401a8daec606a6ed647d |
---|---|
record_format |
Article |
spelling |
doaj-9044e5bb4d9a401a8daec606a6ed647d2020-11-25T03:31:45ZengBMCGenome Biology1474-760X2020-07-0121111610.1186/s13059-020-02103-2Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomicsKwangbom Choi0Yang Chen1Daniel A. Skelly2Gary A. Churchill3The Jackson LaboratoryUniversity of MichiganThe Jackson LaboratoryThe Jackson LaboratoryAbstract Background Single-cell RNA sequencing is a powerful tool for characterizing cellular heterogeneity in gene expression. However, high variability and a large number of zero counts present challenges for analysis and interpretation. There is substantial controversy over the origins and proper treatment of zeros and no consensus on whether zero-inflated count distributions are necessary or even useful. While some studies assume the existence of zero inflation due to technical artifacts and attempt to impute the missing information, other recent studies argue that there is no zero inflation in scRNA-seq data. Results We apply a Bayesian model selection approach to unambiguously demonstrate zero inflation in multiple biologically realistic scRNA-seq datasets. We show that the primary causes of zero inflation are not technical but rather biological in nature. We also demonstrate that parameter estimates from the zero-inflated negative binomial distribution are an unreliable indicator of zero inflation. Conclusions Despite the existence of zero inflation in scRNA-seq counts, we recommend the generalized linear model with negative binomial count distribution, not zero-inflated, as a suitable reference model for scRNA-seq analysis.http://link.springer.com/article/10.1186/s13059-020-02103-2Single-cell RNA sequencingZero inflationBayesian model selectionCell heterogeneityGene expression stochasticity |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Kwangbom Choi Yang Chen Daniel A. Skelly Gary A. Churchill |
spellingShingle |
Kwangbom Choi Yang Chen Daniel A. Skelly Gary A. Churchill Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics Genome Biology Single-cell RNA sequencing Zero inflation Bayesian model selection Cell heterogeneity Gene expression stochasticity |
author_facet |
Kwangbom Choi Yang Chen Daniel A. Skelly Gary A. Churchill |
author_sort |
Kwangbom Choi |
title |
Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
title_short |
Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
title_full |
Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
title_fullStr |
Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
title_full_unstemmed |
Bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
title_sort |
bayesian model selection reveals biological origins of zero inflation in single-cell transcriptomics |
publisher |
BMC |
series |
Genome Biology |
issn |
1474-760X |
publishDate |
2020-07-01 |
description |
Abstract Background Single-cell RNA sequencing is a powerful tool for characterizing cellular heterogeneity in gene expression. However, high variability and a large number of zero counts present challenges for analysis and interpretation. There is substantial controversy over the origins and proper treatment of zeros and no consensus on whether zero-inflated count distributions are necessary or even useful. While some studies assume the existence of zero inflation due to technical artifacts and attempt to impute the missing information, other recent studies argue that there is no zero inflation in scRNA-seq data. Results We apply a Bayesian model selection approach to unambiguously demonstrate zero inflation in multiple biologically realistic scRNA-seq datasets. We show that the primary causes of zero inflation are not technical but rather biological in nature. We also demonstrate that parameter estimates from the zero-inflated negative binomial distribution are an unreliable indicator of zero inflation. Conclusions Despite the existence of zero inflation in scRNA-seq counts, we recommend the generalized linear model with negative binomial count distribution, not zero-inflated, as a suitable reference model for scRNA-seq analysis. |
topic |
Single-cell RNA sequencing Zero inflation Bayesian model selection Cell heterogeneity Gene expression stochasticity |
url |
http://link.springer.com/article/10.1186/s13059-020-02103-2 |
work_keys_str_mv |
AT kwangbomchoi bayesianmodelselectionrevealsbiologicaloriginsofzeroinflationinsinglecelltranscriptomics AT yangchen bayesianmodelselectionrevealsbiologicaloriginsofzeroinflationinsinglecelltranscriptomics AT danielaskelly bayesianmodelselectionrevealsbiologicaloriginsofzeroinflationinsinglecelltranscriptomics AT garyachurchill bayesianmodelselectionrevealsbiologicaloriginsofzeroinflationinsinglecelltranscriptomics |
_version_ |
1724571889344446464 |