Scywalker: scalable end-to-end data analysis workflow for nanopore single-cell transcriptome sequencing

Abstract

We introducescywalker, an innovative and scalable package developed to comprehensively analyze long-read nanopore sequencing data of full-length single-cell or single-nuclei cDNA. Existing nanopore single-cell data analysis tools showed severe limitations in handling current data sizes. We developed novel scalable methods for cell barcode demultiplexing and single-cell isoform calling and quantification and incorporated these in an easily deployable package. Scywalker streamlines the entire analysis process, from sequenced fragments in FASTQ format to demultiplexed pseudobulk isoform counts, into a single command suitable for execution on either server or cluster. Scywalker includes data quality control, cell type identification, and an interactive report. Assessment of datasets from the human brain, Arabidopsis leaves, and previously benchmarked data from mixed cell lines, demonstrate excellent correlation with short-read analyses at both the cell-barcoding and gene quantification levels. At the isoform level, we show that scywalker facilitates the direct identification of cell-type-specific expression of novel isoforms.

Advances in single-cell long-read sequencing technologies

“…The scNaST set of tools builds up on scNapBar ( 39 ), using a hybrid approach to assign spatial barcodes to long reads, followed by spatial spot deconvolution, spatial gene expression, transcript classification, and differential transcript usage ( 47 ). Only recently, two toolkits, scNanoGPS (single-cell Nanopore sequencing analysis of Genotypes and Phenotypes Simultaneously) and Scywalker (not peer-reviewed yet), were developed to assign CBs and UMIs as well as calculate gene and transcript-wise expression profiles without complementary single-cell short reads ( 67 , 68 ). scNanoGPS features CB correction that does not rely on a user-provided barcode whitelist; instead, it is generated within the algorithm.…”