Sockeye: nanopore-only demultiplexing of single-cell reads

Abstract

The combination of single-cell isolation methods, novel barcoding schemes, and sequencing-based readouts has led to an explosion in the number of single-cell genomic assays available. These assays can be combined to create a multi-modal snapshot of cell state, providing unprecedented insight into the heterogeneity associated with different cell and tissue types, developmental trajectories, and different disease states. Multiple studies have now shown that nanopore sequencing of 10x libraries can be used in tandem with short-read sequencing to provide a more complete view of the transcriptome, revealing transcript isoforms and associated genotypes inaccessible to short reads alone. Several tools have been developed to analyse nanopore-sequenced 10x transcriptome libraries; however, they currently assume access to paired short-read data. Given the throughput and accuracy improvements of the Oxford Nanopore platform, this is no longer necessary for many analyses. Here, we present Sockeye — a bioinformatic pipeline that identifies 10x cell barcodes and UMIs without the use of short-read data, outputting files compatible with common downstream analyses.