User-friendly, scalable tools and workflows for single-cell RNAseq analysis

To the Editor — As Single-cell RNA-Seq (scRNA-Seq) becomes widespread, accessible and scalable computational pipelines for data analysis are needed. We introduce an interactive computational environment for single cell studies, based on Galaxy¹, with functions from established workflows. Single Cell interactive application (SCiAp) provides easy access to data from the Human Cell Atlas (HCA) and EMBL-EBI’s Single Cell Expression Atlas (SCEA)² projects and can be deployed in different computing platforms, making single-cell data analysis of large-scale projects accessible to the scientific community.

Consortia such as the HCA, the Fly Cell Atlas, among others, are generating large amounts of scRNA-Seq datasets that will be available for researchers to re-utilize and to analyze their own datasets. For instance, SCEA provides scRNA-Seq datasets comprising >3 million cells, across cell types, tissues and more than14 species. This large collection of scRNA-Seq data demands adequate computational infrastructure, analysis tools and workflows to help researchers make the most of it.

The Galaxy framework has enabled flexible and scalable deployment across multiple clouds through the Galaxy-Kubernetes integration³, supporting analysis of large datasets. Galaxy offers a user-friendly framework for building and sharing workflows. It is supported by a vibrant community of bioinformaticians who continuously enrich the tool repository with analysis methods for e.g., scRNA-seq⁴. Built on Galaxy, SCiAp facilitates data access (HCA, SCEA and own data), downstream analysis and visualisation of scRNA-Seq datasets. We share tools and workflows (including those used in SCEA) in SCiAp that can run through the web interface or the command line. An instance, known as the HCA Galaxy instance, is available at https://humancellatlas.usegalaxy.eu/ (Figure 1). Additional technical details and usability among many other topics are covered in the Supplementary Methods.

Figure 1:

(A) Load matrix data from HCA or SCEA directly into SCiAp Galaxy. (B) Run configurable scRNA-Seq analysis through SCiAp. (C) Inspect results interactively through UCSC-CellBrowser and plots within Galaxy.

A key feature of SCiAp is the ability to integrate tools from different workflows, written in different languages. We break monolithic tools into analysis modules, enabling users to try different competing tool-sets and, where possible, integrate them into the same workflows. For example, we produced >20 modules for Scanpy⁵, covering data input, filtering, normalisation, variable genes, clustering, dimensionality reductions and trajectory methods among others. Supplementary Table 1 shows all the tools integrated and the different functional modules in which they were broken; Supplementary Note 1 shows the integration of modules from different tools on analysis workflows. SCiAp provides functionality from Scanpy, Seurat⁶, Monocle3⁷, SC3⁸, SCmap⁹, Scater¹⁰, SCCAF¹¹, SCPred¹², SCEasy and UCSC CellBrowser. Supplementary Figure 1 shows a map of scRNA-seq data analysis functionalities that are covered by tool wrappers contributed as part of this work and external contributions incorporated, shown accordingly.

In summary, SCiAp is a suite of components derived from commonly used tools in scRNA-Seq analysis. Being based on Galaxy, it can be deployed on large computational infrastructures or on existing Galaxy instances, reducing software engineering complexities for the biological research community. Supplementary Table 2 shows an overview comparison between SCiAp and similar services. SCiAp outperforms for wider accessibility and the variety of tool sets provided. We also provide the underlying tools that resolve software dependencies via Bioconda¹³ and Biocontainers¹⁴, commonly used frameworks in bioinformatics. Lab-based scientists with a deep understanding of a cellular system can use this computational framework to interrogate scRNA-Seq data, propose further hypotheses and guide their experiments to explore the translational potential of large-scale, single cell studies using the friendly Galaxy environment.