Imperial College London
Portrait Picture

DrNathanSkene

Faculty of Medicine, Department of Brain Sciences

Lecturer in Dementia Research, UK DRI Group Leader
 
 
 
//

Contact

 

n.skene Website

 
 
//

Location

 

515Burlington DanesHammersmith Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Murphy:2022:10.1101/2022.02.16.480517,
author = {Murphy, AE and Skene, NG},
doi = {10.1101/2022.02.16.480517},
title = {A balanced measure shows superior performance of pseudobulk methods over mixed models and pseudoreplication approaches in single-cell RNA-sequencing analysis},
url = {http://dx.doi.org/10.1101/2022.02.16.480517},
year = {2022}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - <jats:title>Summary</jats:title><jats:p>Recently, Zimmerman<jats:italic>et al</jats:italic>.,<jats:sup>1</jats:sup>highlighted the importance of accounting for the dependence between cells from the same individual when conducting differential expression analysis on single-cell RNA-sequencing data. Their work proved the inadequacy of pseudoreplication approaches for such analysis – This was an important step forward that was conclusively proven by them. A hierarchical single-cell expression simulation approach (<jats:underline>hierarchicell</jats:underline>) was developed by Zimmerman<jats:italic>et al</jats:italic>.,<jats:sup>1</jats:sup>to generate non-differentially expressed genes upon which performance was evaluated using the type 1 error rate; the proportion of non-differentially expressed genes indicated as differentially expressed by a model. However, evaluating such models on their type 1 or type 2 error rate in isolation is insufficient to determine their true performance – for example, a method with low type 1 error may have a high type 2 error rate. Moreover, because no seed was set for the pseudo-random number generator used in hierarchicell, the different methods evaluated by Zimmerman<jats:italic>et al</jats:italic>. were done so on different simulated datasets. Here, we corrected these issues, reran the author’s analysis and found pseudobulk methods outperformed mixed models.</jats:p><jats:sec><jats:title>Contact</jats:title><jats:p>Alan Murphy:<jats:email>a.murphy@imperial.ac.uk</jats:email>, Nathan Skene:<jats:email>n.skene@imperial.ac.uk</jats:email></jats:p></jats:sec><jats:sec><jats:title>Code availability</jats:title><jats:p>The modified version of hierarchicell which returns all error metrics, uses the same simulated data across approaches and has ch
AU  - Murphy,AE
AU  - Skene,NG
DO  - 10.1101/2022.02.16.480517
PY  - 2022///
TI  - A balanced measure shows superior performance of pseudobulk methods over mixed models and pseudoreplication approaches in single-cell RNA-sequencing analysis
UR  - http://dx.doi.org/10.1101/2022.02.16.480517
ER  -
Download