Regulatory network-based imputation of dropouts in single-cell RNA sequencing data

Single-cell RNA sequencing (scRNA-seq) methods are typically unable to quantify the expression levels of all genes in a cell, creating a need for the computational prediction of missing values (‘dropout imputation’). Most existing dropout imputation methods are limited in the sense that they exclusively use the scRNA-seq dataset at hand and do not exploit external gene-gene relationship information. Here, we show that a transcriptional regulatory network learned from external, independent gene expression data improves dropout imputation. Using a variety of human scRNA-seq datasets we demonstrate that our network-based approach outperforms published state-of-the-art methods. The network-based approach performs particularly well for lowly expressed genes, including cell-type-specific transcriptional regulators. Additionally, we tested a baseline approach, where we imputed missing values using the sample-wide average expression of a gene. Unexpectedly, up to 48% of the genes were better predicted using this baseline approach, suggesting negligible cell-to-cell variation of expression levels for many genes. Our work shows that there is no single best imputation method; rather, the best method depends on gene-specific features, such as expression level and expression variation across cells. We thus implemented an R-package called ADImpute (available from https://github.com/anacarolinaleote/ADImpute) that automatically determines the best imputation method for each gene in a dataset.