CIDR: Ultrafast and accurate clustering through imputation for single-cell RNA-seq data

By Peijie Lin, Michael Troup, Joshua Wing Kei Ho. Genome Biology, 2016

501 citations recorded by the source corpus.

Most existing dimensionality reduction and clustering packages for single-cell RNA-seq (scRNA-seq) data deal with dropouts by heavy modeling and computational machinery. Here, we introduce CIDR (Clustering through Imputation and Dimensionality Reduction), an ultrafast algorithm that uses a novel yet very simple implicit imputation approach to alleviate the impact of dropouts in scRNA-seq data in a principled manner. Using a range of simulated and real data, we show that CIDR improves the standard principal component analysis and outperforms the state-of-the-art methods, namely t-SNE, ZIFA, and RaceID, in terms of clustering accuracy. CIDR typically completes within seconds when processing a data set of hundreds of cells and minutes for a data set of thousands of cells. CIDR can be downloaded at https://github.com/VCCRI/CIDR.

Subfields: Molecular Biology

View Paper: https://www.semanticscholar.org/paper/d1dc3420d200bd6429f85084e63bc4b70b8d0d42

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Eleven grand challenges in single-cell data science (2020, methodology): “[2018] Lin_DAE denoising AE (imputation across k-nearest neighbor genes) Lin et al. [2017a] SAUCIE AE (MMD loss function) Amodio et al. [2019] scScope iterative AE Deng et al. [2019] scVAE Gaussian-mixture VAE (NB / ZINB / ZIP model) Grønbech et al. [2019] scVI VAE (ZINB model) Lopez et al. [2018] s”
Clustering single-cell RNA-seq data with a model-based deep learning approach (2019, methodology): “…tools have been developed recently in this regard, including statistical models CIDR (clustering through imputation and dimensionality reduction) 18 , scImpute 19 , MAGIC 20 and SAVER 21 , to name a few, and deep learning approaches DeepImpute 22 , DCA (deep count autoencoder) 23 and scScope 24…”
An accurate and robust imputation method scImpute for single-cell RNA-seq data (2018, methodology): “An attractive advantage of scImpute is that it can be incorporated into most existing pipelines or downstream analysis of scRNA-seq data, such as normalization 4, 37 , differential expression analysis 7, 32 , clustering and classi ﬁ cation 12, 13 , etc.”
Splatter: simulation of single-cell RNA sequencing data (2017, methodology): “This approach is taken by tools such as SC3 [5], CIDR [6] and Seurat [7], and is appropriate for a sample of mature cells where it is reasonable to expect cells to have a particular type.”
DrImpute: imputing dropout events in single cell RNA sequencing data (2017, methodology): “We hypothesized that 1) 150 preprocessing the scRNA-seq data by imputing the dropout events via DrImpute 151 will improve the accuracy of these clustering methods and 2) the performance of 152 the existing tools combined with DrImpute will perform better than with CIDR or 153 scImpute in addressing ”