nf-core/marsseq: systematic preprocessing pipeline for MARS-seq experiments

Abstract Motivation Single sequencing technology (scRNA-seq) enables the study of gene regulation at a single cell level. Although many sc-RNA-seq protocols have been established, they have varied in technical complexity, sequencing depth and multimodal capabilities leading to shared limitations in data interpretation due to a lack of standardized preprocessing and consistent data reproducibility. While plate based techniques such as Massively Parallel RNA Single cell Sequencing (MARS-seq2.0) provide reference data on the cells that will be sequenced, the data format limits the possible analysis. Here, we focus on the standardization of MARS-seq analysis and its applicability to RNA velocity. Results We have taken the original MARS-seq2.0 pipeline and revised it to enable implementation using the nf-core framework. By doing so, we have simplified pipeline execution, enabling a streamlined application with increased transparency and scalability. We have incorporated additional checkpoints to verify experimental metadata and improved the pipeline by implementing a custom workflow for RNA velocity estimation. The pipeline is part of the nf-core bioinformatics community and is freely available at https://github.com/nfcore/marsseq with data analysis at https://github.com/brickmanlab/proks-et-al-2023. Availability and implementation We introduce an updated preprocessing pipeline for MARS-seq experiments following state-of-the-art guidelines for scientific software development with the added ability to infer RNA velocity.