EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data

Lun, Aaron T. L.; Riesenfeld, Samantha J.; Andrews, Tallulah; Gomes, Tomás; Marioni, John C.

doi:10.1186/s13059-019-1662-y

Cited by 674 publications

(566 citation statements)

References 28 publications

Supporting

Mentioning

564

Contrasting

Order By: Relevance

“…Cell-free RNA is present in the aqueous cell suspension, either as a result of normal biological processes or as a result of tissue dissociation, cell death, or other stresses experienced by cells during the isolation protocol which may cause cells to die or lyse. Such a mechanism has been proposed by others as well [1,4,10].…”

Section: Exhibitmentioning

confidence: 56%

“…A variety of heuristics are typically employed in order determine cutoffs for thresholding cells versus empty droplets. More principled approaches have recently been developed, including EmptyDrops [10], which uses statistical tests to ascertain which droplets have expression profiles significantly different from empty droplets, and DropEst [20], which distinguishes empty and non-empty droplets using a linear classifier trained on features extracted from a set of a priori known empty and non-empty droplets. These approaches, however, depend on having prior knowledge of a range of cell-free droplets (e.g.…”

Section: Accurate Detection Of Empty Dropletsmentioning

confidence: 99%

“…As suggested by previous authors [10,20], other criteria can be used to post-filter cells, including mitochondrial read fraction. Fig.…”

Section: Accurate Detection Of Empty Dropletsmentioning

confidence: 99%

See 2 more Smart Citations

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`

Fleming

Marioni

Babadi

2019

Preprint

Self Cite

193

166

View full text Add to dashboard Cite

Droplet-based scRNA-seq assays are known to produce a significant amount of background RNA counts, the hallmark of which is non-zero transcript counts in presumably empty droplets. The presence of background RNA can lead to systematic biases and batch effects in various downstream analyses such as differential expression and marker gene discovery. In this paper, we explore the phenomenology and mechanisms of background RNA generation in dropletbased scRNA-seq assays and present a deep generative model of background-contaminated counts mirroring those mechanisms. The model is used for learning the background RNA profile, distinguishing cell-containing droplets from empty ones, and retrieving background-free gene expression profiles. We implement the model along with a fast and scalable inference algorithm as the remove-background module in CellBender, an open-source scRNA-seq data processing software package. Finally, we present simulations and investigations of several scRNA-seq datasets to show that processing raw data using CellBender significantly boosts the magnitude and specificity of differential expression across different cell types.

show abstract

Section: Exhibitmentioning

confidence: 56%

Section: Accurate Detection Of Empty Dropletsmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`

Fleming

Marioni

Babadi

2019

Preprint

Self Cite

193

166

View full text Add to dashboard Cite

show abstract

“…Alignment, read filtering, barcode and UMI counting were performed using kallisto-bustools 12 . High quality barcodes were selected based on the overall UMI distribution using emptyDrops 13 . All further analyses were run using the Python-based Scanpy 14 .…”

Section: Bioinformatic Pipelinesmentioning

confidence: 99%

Resolving cellular systems by ultra-sensitive and economical single-cell transcriptome filtering

Vallejo

Davies

Grover

et al. 2019

Preprint

View full text Add to dashboard Cite

Single-cell transcriptomics has sensitivity limits that restrict low abundance transcript identification, affects clustering and introduce artefact. Here, we describe Constellation DropSeq (C-DropSeq), a molecular transcriptome filter that delivers two orders of magnitude sensitivity gains by maximising read utility while reducing sequencing depth and costs. The simple and powerful method is broadly compatible with library preparation routines and was demonstrated by identifying and characterizing the activation of rare dendritic cell sub-populations. MainThe dramatic uptake and expansion of single-cell transcriptome analysis tools has transformed biological research, enabling reconstruction of population architectures and underlying processes to be revealed. The tools rely on compartmentalisation of single cells with the introduction of unique genetic barcodes during library preparation 1 . Though formidable, not unexpectedly these methods have sensitivity limits, with associated transcript absence events (dropouts) that restrict the faithful delineation of cell subtypes and especially overlook low abundant transcripts such as transcription factors, receptors and signalling molecules that are often pivotal for accurately describing cell processes and fate 2,3 . This is a consequence of high abundance transcripts occupying the available NGS read space and exacerbated by exponential PCR-directed library preparation routines.Targeted approaches forgo global transcriptome screens, preferring to select transcripts of known utility and are especially favoured for mechanistic studies. Diverse targeted strategies have emerged; physical recovery of transcriptome subsets 4 , coupling custom primers to poly(dT) capture beads (DART-seq) 5 and panel selection by PCR as with the Rhapsody workflow (BD) 6 . These methods are technically challenging and introduce substantial costs. Here, we describe Constellation DropSeq (C-DropSeq), a remarkably simple, inexpensive and scalable (e.g. >200 targets) approach, introducing a linear amplification stage in advance of conventional library preparation. Superior performance is

show abstract

“…For example, the cellRanger-atac method fits a mixture of two zero-inflated negative binomial models to discriminate cell barcodes and non-cell barcodes. EmptyDrops [41], originally designed to identify cells from scRNA-Seq data, models the counts using a Dirichlet-multinomial distribution. scATAC-pro provides all of the aforementioned strategies/methods.…”

Section: Cell Callingmentioning

confidence: 99%

scATAC-pro: a comprehensive workbench for single-cell chromatin accessibility sequencing data

Yu¹,

Zhu

Chen³

et al. 2019

Preprint

View full text Add to dashboard Cite

AbstractSingle cell chromatin accessibility sequencing (scCAS) has become a powerful technology for understanding epigenetic heterogeneity of complex tissues. The development of several experimental protocols has led to a rapid accumulation of scCAS data. In contrast, there is a lack of open-source software tools for comprehensive processing, analysis and visualization of scCAS data generated using all existing experimental protocols. Here we present scATAC-pro for quality assessment, analysis and visualization of scCAS data. scATAC-pro provides flexible choice of methods for different data processing and analytical tasks, with carefully curated default parameters. A range of quality control metrics are computed for several key steps of the experimental protocol. scATAC-pro generates summary reports for both quality assessment and downstream analysis. It also provides additional utility functions for generating input files for various types of downstream analyses and data visualization. With the rapid accumulation of scCAS data, scATAC-pro will facilitate studies of epigenomic heterogeneity in healthy and diseased tissues.

show abstract

EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data

Cited by 674 publications

References 28 publications

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`

Resolving cellular systems by ultra-sensitive and economical single-cell transcriptome filtering

scATAC-pro: a comprehensive workbench for single-cell chromatin accessibility sequencing data

Contact Info

Product

Resources

About

EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data

Cited by 674 publications

References 28 publications

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using CellBender

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using CellBender

Resolving cellular systems by ultra-sensitive and economical single-cell transcriptome filtering

scATAC-pro: a comprehensive workbench for single-cell chromatin accessibility sequencing data

Contact Info

Product

Resources

About

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`

Unsupervised removal of systematic background noise from droplet-based single-cell experiments using `CellBender`