Single-cell transcriptomics suffer from sensitivity limits that restrict low abundance transcript identification, affects clustering and can hamper downstream analyses. Here, we describe Constellation sequencing (Constellation-Seq), a molecular transcriptome filter that delivers two orders of magnitude sensitivity gains by maximizing read utility while reducing the data sparsity and sequencing costs. The technique reliably measures changes in gene expression and was demonstrated by resolving rare dendritic cell populations from a peripheral blood mononuclear cell sample sample and exploring their biology with extreme resolution. The simple and powerful method is fully compatible with standard scRNA-Seq library preparation protocols and can be used for hypothesis testing, marker validation or investigating pathways.
