Granatum: a graphical single-cell RNA-Seq analysis pipeline for genomics scientists

Abstract

Background: Single-cell RNA sequencing (scRNA-Seq) is an increasingly popular platform to study heterogeneity at the single-cell level. Computational methods to process scRNA-Seq data are not very accessible to bench scientists as they require a significant amount of bioinformatic skills.

Results: We have developed Granatum, a web-based scRNA-Seq analysis pipeline to make analysis more broadly accessible to researchers. Without a single line of programming code, users can click through the pipeline, setting parameters and visualizing results via the interactive graphical interface. Granatum conveniently walks users through various steps of scRNA-Seq analysis. It has a comprehensive list of modules, including plate merging and batch-effect removal, outlier-sample removal, gene-expression normalization, imputation, gene filtering, cell clustering, differential gene expression analysis, pathway/ontology enrichment analysis, protein network interaction visualization, and pseudo-time cell series construction.

Conclusions: Granatum enables broad adoption of scRNA-Seq technology by empowering bench scientists with an easy-to-use graphical interface for scRNA-Seq data analysis. The package is freely available for research use at http://garmiregroup.org/granatum/app.

Keywords: Clustering; Differential expression; Gene expression; Graphical; Imputation; Normalization; Pathway; Pseudo-time; Single-cell; Software.

Fig. 1

Granatum workflow. Granatum is built with the Shiny framework, which integrates the front-end with the back-end. A public server has been provided for easy access, and local deployment is also possible. The user uploads one or more expression matrices with corresponding metadata for samples. The back-end stores data separately for each individual user, and invokes third-party libraries on demand

Fig. 2

Batch-effect removal. The PCA plots show the before/after median alignment comparison. The colors indicate the two batches 1 and 2, and the shapes indicate the three cell types reported from the original data. a Before batch-effect removal; b after batch-effect removal

Fig. 3

Outlier removal using PCA plot. a Before outlier removal. b After outlier removal

Fig. 4

Box-plot comparison of normalization methods. The cell size is down-sampled to representatively show the general effect of each method. The colors indicate the three cell types reported from the original data. a Original data (no normalization). b Quantile normalization. c Geometrical mean normalization. d Size-factor normalization. e Voom normalization

Fig. 5

Comparison of DE genes identified by Granatum or ASAP pipeline. a MA plot. Blue color labels DE genes, and gray dots are non-DE genes. b Venn diagram showing the number of DE genes identified by both methods, as well as those uniquely identified by either pipeline. c Bar chart comparing the number of genes up regulated in primary cells (red) or metastasized cells (green). d Bubble plots of KEGG pathway GSEA results for the DE genes identified by either pipeline. The y-axis represents the enrichment score of the gene sets, the x-axis shows gene set names, and the size of the bubble indicates the number of genes in that gene set

Fig. 6

Protein–protein interaction network and pseudo-time construction steps. a The PPI network derived from the DE results between PDX primary and metastasized cells in the K-dataset. The color on each node (gene) indicates its Z-score in the differential expression test. Red and blue indicate up- and down-regulation in metastasized cells, respectively. b The pseudo-time construction step. The Monocle algorithm is customized to visualize the paths among individual cells. Sample labels from the metadata are shown as different colors in the plot