scRMD: imputation for single cell RNA-seq data via robust matrix decomposition

Abstract

Motivation Single cell RNA-sequencing (scRNA-seq) technology enables whole transcriptome profiling at single cell resolution and holds great promises in many biological and medical applications. Nevertheless, scRNA-seq often fails to capture expressed genes, leading to the prominent dropout problem. These dropouts cause many problems in down-stream analysis, such as significant increase of noises, power loss in differential expression analysis and obscuring of gene-to-gene or cell-to-cell relationship. Imputation of these dropout values can be beneficial in scRNA-seq data analysis. Results In this article, we model the dropout imputation problem as robust matrix decomposition. This model has minimal assumptions and allows us to develop a computational efficient imputation method called scRMD. Extensive data analysis shows that scRMD can accurately recover the dropout values and help to improve downstream analysis such as differential expression analysis and clustering analysis. Availability and implementation The R package scRMD is available at https://github.com/XiDsLab/scRMD. Supplementary information Supplementary data are available at Bioinformatics online.

SERGIO: A Single-Cell Expression Simulator Guided by Gene Regulatory Networks

“…Simulation tools (''simulators'') for single-cell expression data have been reported in various forms. Several studies offering novel analysis tools use in-house simulators to benchmark those tools (Van den Berge et al, 2018;Campbell and Yau, 2018;Chen et al, 2020;Gong et al, 2019;Korthauer et al, 2016;Risso et al, 2018;Wolf et al, 2018), while other studies specifically develop simulators for use by the community (Holm, 2019;Marouf et al, 2020;Papadopoulos et al, 2019;Vieth et al, 2017;Zappia et al, 2017;Zhang et al, 2019). Most of these simulators are geared toward capturing the noise characteristics of technologies such as single-cell RNA-seq (scRNA-seq), by first estimating statistical quantities describing real datasets and then sampling singlecell expression profiles from probability distributions that mirror those quantities.…”

RobustClone: A robust PCA method of tumor clone and evolution inference from single-cell sequencing data

“…The RPCA-based algorithm was applied to the scSMD method for the imputation of singlecell RNA sequencing data (35). Their model assumes that dropout events should be relatively sparse in the original gene expression matrix (35). In the case of the single-cell DNA genotype matrix, however, missing entries may reach as high as 58% (22).…”

Correlation Imputation in Single cell RNA-seq using Auxiliary Information and Ensemble Learning

“…Based on the similarity among cells' gene expressions, both drImpute [8] and PRIME [14] impute the dropouts of a cell by using the gene expressions of the cells belonging to the same cluster. The scRMD methodology [1] infers the gene expressions of cells by robust matrix decomposition.…”

“…Unfortunately, the presence of dropouts can cause missingness in the sample covariance matrix. For this last problem there are plenty of covariance matrix completion methodologies that may be used [9,22], but these, just like the data imputation methods [1,8,12,14], perform ideally under different assumptions about the data.…”