Ranu, N. | bioRxiv Science

bioRxiv2017

Targeting individual cells by barcode in pooled sequence libraries

There is rising interest in applying single-cell transcriptome analysis and other single-cell sequencing methods to resolve differences between cells. Pooled processing of thousands of single cells is now routinely practiced by introducing cell-specific DNA barcodes early in cell processing protocols1-5. However, researchers must sequence a large number of cells to sample rare subpopulations6-8, even when fluorescence-activated cell sorting (FACS) is used to pre-enrich rare cell populations. Here, a new molecular enrichment method is used in conjunction with FACS enrichment to enable efficient sampling of rare dendritic cell (DC) populations, including the recently identified AXL+SIGLEC6+ (AS DCs) subset7, within a 10X Genomics single-cell RNA-Seq library. DC populations collectively represent 1-2% of total peripheral blood mononuclear cells (PBMC), with AS DC representing only 1-3% of human blood DCs and 0.01-0.06% of total PBMCs.

bioengineering↗

bioRxiv2017

Estimation of immune cell content in tumour tissue using single-cell RNA-seq data

As interactions between the immune system and tumour cells are governed by a complex network of cell-cell interactions, knowing the specific immune cell composition of a solid tumour may be essential to predict a patients response to immunotherapy. Here, we analyse in depth how to derive the cellular composition of a solid tumour from bulk gene expression data by mathematical deconvolution, using indication- and cell type-specific reference gene expression profiles (RGEPs) from tumour-derived single-cell RNA sequencing data. We demonstrate that tumour-derived RGEPs are essential for the successful deconvolution and that RGEPs from peripheral blood are insufficient. We distinguish nine major cell types as well as three T cell subtypes. As the ratios of CD4+, CD8+ and regulatory T cells have been shown to predict overall survival, we extended our analysis to include the estimation of prognostic ratios that may enable the application in a clinical setting. Using the tumour derived RGEPs, we can estimate, for the first time, the content of cancer associated fibroblasts, endothelial cells and the malignant cells in a patient sample by a deconvolution approach. In addition, improved tumour cell gene expression profiles can be obtained by this method by computationally removing contamination from non-malignant cells. Given the difficulty around sample preparation and storage to obtain high quality single-cell RNA-seq data in the clinical context, the presented method represents a computational solution to derive the cellular composition of a tissue sample.

bioinformatics↗