demuxlet

demuxlet is a computational tool to address the challenges associated with droplet single-cell RNA-sequencing (dscRNA-seq), such as inefficient sample processing and technical batch effects. This innovative tool leverages natural genetic variation to accurately determine the sample identity of each droplet, which contains a single cell (referred to as a singlet), and to detect droplets that contain two cells (doublets). This capability is crucial for enabling multiplexed dscRNA-seq experiments, where cells from unrelated individuals can be pooled together and captured at a higher throughput than is possible with standard workflows.

The utility of demuxlet has been demonstrated through simulations, which show that with just 50 single-nucleotide polymorphisms (SNPs) per cell, it can assign 97% of singlets and identify 92% of doublets in pools comprising up to 64 individuals. This high level of accuracy in singlet assignment and doublet identification significantly enhances the efficiency and reliability of dscRNA-seq experiments.

Further validation of demuxlet’s capabilities comes from its application to real datasets. When given genotyping data for each of the eight pooled samples, demuxlet successfully recovered the sample identity of more than 99% of singlets and identified doublets at rates consistent with previous estimates.