Epi-Impute: Single-Cell RNA-seq Imputation via Integration with Single-Cell ATAC-seq

• Published in: International journal of molecular sciences 2023-03, Vol.24 (7), p.6229
• Main Authors: Raevskiy, Mikhail, Yanvarev, Vladislav, Jung, Sascha, Del Sol, Antonio, Medvedeva, Yulia A
• Format: Article
• Language: English
• Subjects: Analysis; Chromatin Immunoprecipitation Sequencing; Computer applications; Datasets; Dropouts; Error reduction; Gene expression; Gene Expression Profiling; Genes; imputation; Messenger RNA; Methods; Probability; Sequence Analysis, RNA - methods; single cell ATAC-seq; single cell RNA-seq; Single-Cell Analysis - methods; Single-Cell Gene Expression Analysis; Software
• ISSN: 1422-0067; 1661-6596; 1422-0067
• DOI: 10.3390/ijms24076229

Single-cell RNA-seq data contains a lot of dropouts hampering downstream analyses due to the low number and inefficient capture of mRNAs in individual cells. Here, we present Epi-Impute, a computational method for dropout imputation by reconciling expression and epigenomic data. Epi-Impute leverages single-cell ATAC-seq data as an additional source of information about gene activity to reduce the number of dropouts. We demonstrate that Epi-Impute outperforms existing methods, especially for very sparse single-cell RNA-seq data sets, significantly reducing imputation error. At the same time, Epi-Impute accurately captures the primary distribution of gene expression across cells while preserving the gene-gene and cell-cell relationship in the data. Moreover, Epi-Impute allows for the discovery of functionally relevant cell clusters as a result of the increased resolution of scRNA-seq data due to imputation.