Scdrake: a reproducible and scalable pipeline for scRNA-seq data analysis

MOTIVATION: While the workflow for primary analysis of single-cell RNA-seq (scRNA-seq) data is well established, the secondary analysis of the feature-barcode matrix is usually done by custom scripts. There is no fully automated pipeline in the R statistical environment, which would follow the current best programming practices and requirements for reproducibility. RESULTS: We have developed scdrake, a fully automated workflow for secondary analysis of scRNA-seq data, which is fully implemented in the R language and built within the drake framework. The pipeline includes quality control, cell and gene filtering, normalization, detection of highly variable genes, dimensionality reduction, clustering, cell type annotation, detection of marker genes, differential expression analysis and integration of multiple samples. The pipeline is reproducible and scalable, has an efficient execution, provides easy extendability and access to intermediate results and outputs rich HTML reports. Scdrake is distributed as a Docker image, which provides a straightforward setup and enhances reproducibility. AVAILABILITY AND IMPLEMENTATION: The source code and documentation are available under the MIT license at https://github.com/bioinfocz/scdrake and https://bioinfocz.github.io/scdrake, respectively. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics Advances online.