Comparative analysis of differential gene expression analysis tools for single-cell RNA sequencing data

Wang, Tianyu; Li, Boyang; Nelson, Craig E.; Nabavi, Sheida

doi:10.1186/s12859-019-2599-6

Cited by 257 publications

(245 citation statements)

References 47 publications

Supporting

Mentioning

240

Contrasting

Order By: Relevance

“…As noted above, singleCellHaystack returns inflated p-values, because the input coordinates (PCs, t-SNE or UMAP coordinates) are dimensions which contain a large proportion of the variability in the original data. Clustering-based DEG prediction methods appear to suffer from this problem even more, because of their double use of gene expression data (for defining clusters and for DEG prediction) (13,14). In future updates we hope to address this issue.…”

Section: Discussionmentioning

confidence: 99%

“…On top of that, there is an inherent difficulty in deciding a suitable number of clusters in data that is high-dimensional and therefore hard to visualize. Finally, consistency between DEG prediction methods has been reported to be low (14). In contrast, singleCellHaystack works independently of any grouping or clustering of cells.…”

Section: Comparison With Other Methodsmentioning

confidence: 99%

“…This approach for finding differentially expressed genes by comparing between clusters is widely used in existing methods (8,11,12), and enables finding cluster-specific marker genes that facilitate labeling different cell populations. However, recent comparisons found that DEG prediction approaches for bulk RNA-seq do not generally perform worse than methods designed specifically for single-cell RNA-seq (scRNA-seq), and that agreement between existing methods is low (13,14). Defining more flexible statistical frameworks for predicting complex patterns of differential expression is one of the grand challenges in single cell data analysis (15).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

singleCellHaystack: Finding surprising genes in 2-dimensional representations of single cell transcriptome data

Vandenbon

Díez

2019

Preprint

View full text Add to dashboard Cite

A common analysis of single-cell sequencing data includes dimensionality reduction using t-SNE or UMAP, clustering of cells, and identifying differentially expressed genes. How cell clusters are defined has important consequences in the interpretation of results and downstream analyses, but is often not straightforward. To address this difficulty, we present a new approach called singleCellHaystack that enables the identification of differentially expressed genes (DEGs) without relying on explicit clustering of cells. Our method uses Kullback-Leibler Divergence to find genes that are expressed in subsets of cells that are non-randomly positioned in a multidimensional space. We illustrate the usage of singleCellHaystack through applications on several single-cell datasets, demonstrate that it enables the identification of markers important for cell subset separation in an unbiased way, and compare its results with those of a traditional, clustering-based DEG prediction method.Availability and implementation: singleCellHaystack is implemented as an R package and is available from https://github.com/alexisvdb/singleCellHaystack Contact: alexisvdb@infront.kyoto-u.ac.jp

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Comparison With Other Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

singleCellHaystack: Finding surprising genes in 2-dimensional representations of single cell transcriptome data

Vandenbon

Díez

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…One of the more common applications of RNA-seq data is estimating and testing for differences in gene expression between two groups. Many packages and techniques exist to perform this task [Robinson and Smyth, 2007b, Hardcastle and Kelly, 2010, Van De Wiel et al, 2012, Kharchenko et al, 2014, Law et al, 2014, Love et al, 2014, Finak et al, 2015, Guo et al, 2015, Nabavi et al, 2015, Delmans and Hemberg, 2016, Korthauer et al, 2016, Costa-Silva et al, 2017, Qiu et al, 2017, Miao et al, 2018, Van den Berge et al, 2018, Wang and Nabavi, 2018, Wang et al, 2019, and so developing approaches and software to compare these different software packages would be of great utility to the scientific community. Generating data from the two-group model is a special case of (1) and (2), where…”

Section: Application: Evaluating Differential Expression Analysismentioning

confidence: 99%

Data-based RNA-seq Simulations by Binomial Thinning

Gerard

2019

Preprint

View full text Add to dashboard Cite

With the explosion in the number of methods designed to analyze bulk and single-cell RNAseq data, there is a growing need for approaches that assess and compare these methods. The usual technique is to compare methods on data simulated according to some theoretical model. However, as real data often exhibit violations from theoretical models, this can result in unsubstantiated claims of a method's performance. Rather than generate data from a theoretical model, in this paper we develop methods to add signal to real RNA-seq datasets. Since the resulting simulated data are not generated from an unrealistic theoretical model, they exhibit realistic (annoying) attributes of real data. This lets RNA-seq methods developers assess their procedures in non-ideal (model-violating) scenarios. Our procedures may be applied to both single-cell and bulk RNA-seq. We show that our simulation method results in more realistic datasets and can alter the conclusions of a differential expression analysis study. We also demonstrate our approach by comparing various factor analysis techniques on RNA-seq datasets. Our tools are available in the seqgendiff R package on the Comprehensive R Archive Network: https://cran.r-project.org/package=seqgendiff.

show abstract

“…A main task in single cell gene expression data analysis (scRNA-Seq) is selecting marker genes after cell types are identified via unsupervised clustering of single cells. This is usually accomplished by differential expression analysis [1,2]. On the experimental side, detecting a cell type's marker gene by immunostaining or in a FACS experiment is often used as a proxy for identifying all cells comprising that cell type.…”

Section: Introductionmentioning

confidence: 99%

genesorteR: Feature Ranking in Clustered Single Cell Data

Ibrahim

Kramann²

2019

Preprint

View full text Add to dashboard Cite

Marker genes identified in single cell experiments are expected to be highly specific to a certain cell type and highly expressed in that cell type. Detecting a gene by differential expression analysis does not necessarily satisfy those two conditions and is typically computationally expensive for large cell numbers.Here we present genesorteR, an R package that ranks features in single cell data in a manner consistent with the expected definition of marker genes in experimental biology research. We benchmark genesorteR using various data sets and show that it is distinctly more accurate in large single cell data sets compared to other methods. genesorteR is orders of magnitude faster than current implementations of differential expression analysis methods, can operate on data containing millions of cells and is applicable to both single cell RNA-Seq and single cell ATAC-Seq data.genesorteR is available at https://github.com/mahmoudibrahim/genesorteR.

show abstract

Comparative analysis of differential gene expression analysis tools for single-cell RNA sequencing data

Cited by 257 publications

References 47 publications

singleCellHaystack: Finding surprising genes in 2-dimensional representations of single cell transcriptome data

singleCellHaystack: Finding surprising genes in 2-dimensional representations of single cell transcriptome data

Data-based RNA-seq Simulations by Binomial Thinning

genesorteR: Feature Ranking in Clustered Single Cell Data

Contact Info

Product

Resources

About