CHETAH: a selective, hierarchical cell type identification method for single-cell RNA sequencing

Kanter, Jurrian K. de; Lijnzaad, Philip; Candelli, Tito; Margaritis, Thanasis; Holstege, Frank C.P.

doi:10.1093/nar/gkz543

Cited by 212 publications

(130 citation statements)

References 60 publications

Supporting

Mentioning

130

Contrasting

Order By: Relevance

“…The k -means clustering is a key ingredient of commonly applied scRNA-seq clustering methods such as SC3 [18] and Waterfall [25]. The hierarchical clustering is a key ingredient of commonly applied scRNA-seq clustering methods such as CIDR [17] and CHETAH [81]. The Louvain method is also a commonly used clustering method for common single-cell analysis software such as Seurat [16] and Monocle [27, 82].…”

Section: Methodsmentioning

confidence: 99%

Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis

et al. 2019

View full text Add to dashboard Cite

BackgroundDimensionality reduction is an indispensable analytic component for many areas of single-cell RNA sequencing (scRNA-seq) data analysis. Proper dimensionality reduction can allow for effective noise removal and facilitate many downstream analyses that include cell clustering and lineage reconstruction. Unfortunately, despite the critical importance of dimensionality reduction in scRNA-seq analysis and the vast number of dimensionality reduction methods developed for scRNA-seq studies, few comprehensive comparison studies have been performed to evaluate the effectiveness of different dimensionality reduction methods in scRNA-seq.ResultsWe aim to fill this critical knowledge gap by providing a comparative evaluation of a variety of commonly used dimensionality reduction methods for scRNA-seq studies. Specifically, we compare 18 different dimensionality reduction methods on 30 publicly available scRNA-seq datasets that cover a range of sequencing techniques and sample sizes. We evaluate the performance of different dimensionality reduction methods for neighborhood preserving in terms of their ability to recover features of the original expression matrix, and for cell clustering and lineage reconstruction in terms of their accuracy and robustness. We also evaluate the computational scalability of different dimensionality reduction methods by recording their computational cost.ConclusionsBased on the comprehensive evaluation results, we provide important guidelines for choosing dimensionality reduction methods for scRNA-seq data analysis. We also provide all analysis scripts used in the present study at www.xzlab.org/reproduce.html.

show abstract

Section: Methodsmentioning

confidence: 99%

Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Our unsupervised approach outperformed CellAssign, a marker gene-based probabilistic cell-type assignment method, which was recently shown to enable accurate annotation of multiple cell types 29 . Another group of supervised methods, such as CaSTLe 31 , ACTINN 32 , SingleR 34 and CHETAH 35 , utilize reference bulk or single-cell transcriptomic data for cell type predictions, and therefore require comprehensive, manually-annotated and high-quality reference datasets; furthermore, these tools do not allow identification of novel cell-type marker genes. In contrast, ScType requires neither reference scRNA-seq datasets nor manual selection of marker genes;…”

Section: Discussionmentioning

confidence: 99%

Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data

Ianevski

Giri

Aittokallio

2019

Preprint

View full text Add to dashboard Cite

Single-cell transcriptomics enables systematic charting of cellular composition of complex tissues.Identification of cell populations often relies on unsupervised clustering of cells based on the similarity of the scRNA-seq profiles, followed by manual annotation of cell clusters using established marker genes. However, manual selection of marker genes for cell-type annotation is a laborious and error-prone task since the selected markers must be specific both to the individual cell clusters and various cell types. Here, we developed a computational method, termed ScType, which enables data-driven selection of marker genes based solely on given scRNA-seq data.Using a compendium of 7 scRNA-seq datasets from various human and mouse tissues, we demonstrate how ScType enables unbiased, accurate and fully-automated single-cell type annotation by guaranteeing the specificity of marker genes both across cell clusters and cell types.The widely-applicable method is implemented as an interactive web-tool (https://sctype.fimm.fi), connected with comprehensive database of specific markers.

show abstract

“…CHEETAH v1.0.4 R package 37 was applied for the comparison of our single nucleus data to other reference datasets. A reference for healthy brain cell types was acquired from the study of Darmanis et al 38 while the reference data set for the tumor cells (H3 K27M mutated pediatric high grade glioma) was from the study of Filbin et al 7 .…”

Section: Methodsmentioning

confidence: 99%

Establishment of a simplified preparation method for single-nucleus RNA-sequencing and its application to long-term frozen tumor tissues

Ernst

Okonechnikov

Bageritz

et al. 2020

Preprint

View full text Add to dashboard Cite

Recent advances allowing the genomic analysis of individual cells from a bulk population have provided intriguing new insights into areas such as developmental processes and tumor heterogeneity. Most approaches to date, however, rely on the availability of fresh surgical specimens, thereby dramatically reducing the ability to profile particularly rare tissue types. Pediatric central nervous system tumors - the leading cause of childhood cancer deaths - represent one such example, where often only frozen rather than native material is available. Due to an increasing need for advanced techniques to understand the heterogeneity of these tumors, we optimized a method to isolate intact nuclei from long-term frozen pediatric glioma tissues. We performed a technical comparison between different single nucleus RNA-sequencing (snRNA-seq) systems using a patient-derived xenograft model as a test sample. Further, we applied the established nucleus isolation method to analyze frozen primary tissue from two pediatric central nervous system tumors - one pilocytic astrocytoma and one glioblastoma - allowing the identification of distinct tumor cell populations and infiltrating microglia. The results show that our fast, simple and low-cost nuclear isolation protocol provides intact nuclei, which can be used in both droplet-based 3' transcriptome amplification (10X Genomics) and plate-based whole transcriptome amplification (Fluidigm C1) single-cell sequencing platforms, thereby dramatically increasing the potential for application of such methods to rare entities.

show abstract

CHETAH: a selective, hierarchical cell type identification method for single-cell RNA sequencing

Cited by 212 publications

References 60 publications

Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis

Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis

Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data

Establishment of a simplified preparation method for single-nucleus RNA-sequencing and its application to long-term frozen tumor tissues

Contact Info

Product

Resources

About