High throughput error correction using dual nucleotide dimer blocks allows direct single-cell nanopore transcriptome sequencing

Droplet-based single-cell sequencing techniques have provided unprecedented insight into cellular heterogeneities within tissues. However, these approaches only allow for the measurement of the distal parts of a transcript following short-read sequencing. Therefore, splicing and sequence diversity information is lost for the majority of the transcript. The application of long-read Nanopore sequencing to droplet-based methods is challenging because of the low base-calling accuracy currently associated with Nanopore sequencing. Although several approaches that use additional short-read sequencing to error-correct the barcode and UMI sequences have been developed, these techniques are limited by the requirement to sequence a library using both short- and long-read sequencing. Here we introduce a novel approach termed single-cell Barcode UMI Correction sequencing (scBUC-seq) to efficiently error-correct barcode and UMI oligonucleotide sequences synthesized by using blocks of dimeric nucleotides. The method can be applied to correct both short-read and long-read sequencing, thereby allowing users to recover more reads per cell that permits direct single-cell Nanopore sequencing for the first time. We illustrate our method by using species-mixing experiments to evaluate barcode assignment accuracy and multiple myeloma cell lines to evaluate differential isoform usage and Ewing's sarcoma cells to demonstrate Ig fusion transcript analysis. Overall design: Mouse (3T3) and human (HEK293T) cells were mixed at a ratio of 1:1 and individual cells were encapsulated using a droplet based single-cell sequencing encapsulator device (DolomiteBio Nadia system). Libraries were then sequenced using a MinION Oxford nanopore flow cell (FLO-MIN106D) or PrimethION flow cell (FLO-PRO002).