Identifying developmentally important genes with single-cell RNA-seq from an embryo

Gene expression studies have typically focused on finding differentially expressed genes or pathways between two or more conditions. More recently, single-cell RNA-seq has been established as a reliable and accessible technique enabling new types of analyses, such as the study of gene expression variation within cell types from cell lines or from relatively similar cells in tissues, organs or tumors. However, although single-cell RNA-seq provides quantitative and comprehensive expression data in a developing embryo, it is not yet clear whether this can replace conventional in situ screens for finding developmentally important genes; moreover, current single-cell data analysis approaches typically cluster cells into types based on a common set of genes or identify more variable or differentially expressed genes using predefined groups of cells, limiting their use for finding genes with novel expression patterns. Here we present a method that comprehensively finds cell-specific patterns of developmentally important regulators directly from single-cell gene expression data of the Ciona embryo, a marine chordate. We recover many of the known expression patterns directly from our single-cell RNA-seq data and despite extensive previous screens, we succeed in finding new cell-specific patterns and genes, which we validate by in situ and single-cell qPCR.