Abstract:
|
Cell type composition of intact bulk tissue can vary across samples. Deciphering cell type composition and its changes during disease progression is an important step towards understanding disease pathogenesis. To infer cell type composition, existing cell type deconvolution methods for bulk RNA-seq data require matched single-cell RNA-seq (scRNA-seq) data, generated under similar clinical conditions, as reference. However, due to the difficulty of obtaining scRNA-seq data in disease samples, only limited scRNA-seq data in matched disease conditions are available. To overcome this limitation, we propose an iterative estimation procedure, MuSiC2, to perform deconvolution analysis of bulk RNA-seq data from different conditions using scRNA-seq data from only one condition as the reference. Benchmark evaluations indicate that MuSiC2 achieves an accurate composition estimation of bulk RNA-seq samples under different conditions, whereas traditional deconvolution methods yield biased cell type composition estimates. We believe the condition-specific cell type composition estimates from MuSiC will facilitate downstream analysis and help identify cellular targets of human diseases.
|