michael-kotliar/cellxgene_VIP

Enables cellxgene to generate violin, stacked violin, stacked bar, heatmap, volcano, embedding, dot, track, density, 2D density, sankey and dual-gene plot in high-resolution SVG/PNG format. It also performs differential gene expression analysis and provides a Command Line Interface (CLI) for advanced users to perform analysis using python and R.

cellxgene VIP unleashes full power of interactive visualization, plotting and analysis of scRNA-seq data in the scale of millions of cells

Since the first single-cell RNA sequencing (scRNA-seq) study was debuted in 2009[1], over 1050 scRNA-seq studies have been published to date. At least 25 studies reported profiles in 200k or more cells[2]. The largest scRNA-seq study reported 2.5 million mouse cells[3]. It is foreseen that there will be a trend of increasing cell size of 500k or more in scRNA-seq studies. The sheer amount of data had brought a challenge in visualizing and interactively exploring such big data set for scientists, even computational savvy specialists.

Cellxgene[4] is a leading open source scRNA-Seq data visualization tool recommended in a recent evaluation[5], which scales well in millions of cells and scores high in user experience by leveraging modern web techniques with interactive features. We tested with a large community of biologists, cellxgene works the best in meeting the need of handling data themselves except lack of some essential plotting seen in scRNA-seq related publications and analysis functions that biologists are used to, hinders its utility and limits scientists from taking advantage of ever accumulating scRNA-seq data in the public to its full potential.

To fill the gap, we developed a plugin of cellxgene named Visualization in Plugin, in short VIP, to address the urgent needs of such essential functions for interactive visual exploration and generation of publication-ready figures. Notably, it enhanced plotting functions significantly to generate violin, stacked violin, stacked bar, heatmap, volcano, embedding, dot, track, density, 2D density, sankey and dual-gene plot in high-resolution by calling server-side scanpy’s[6] plotting functions and general plotting libraries as illustrated in Figure 1 and Supplementary Tutorial.

Figure 1 | cellxgene VIP serves as an ecosystem of plugins which provide essential functions for publication-ready, interactive visualization, as well as CLI for analytics. (svg files were assembled by figureComposer[7] with zoomable version available at https://bit.ly/2QqdMg3 that is best viewed by Chrome) (a) Multi-tSNE/UMAP plot visually highlights which cells expressing cell markers on selected embedding (UMAP based on harmony batch correction in this example). (b) Dual-gene plot highlights cells express SYT1 and GAD1 (green SYT1 only, red GAD1 only, yellow co-expression of STY1 and GAD1), expression cutoff 2.2. (c) Stacked barplot demonstrates the fraction of each major cell type across each sample (C are Control and MS are MS patients). (d) Trackplot shows expression of lineage marker genes across individual cells in annotated clusters. (e) Violin plot shows the AQP4 gene expression across cell types. (f) Sankey diagram (a.k.a. Riverplot) provides quick and easy way to explore the inter-dependent relationship of variables in the MS snRNAseq dataset8. (g) Density plots shows expression of marker genes across annotated clusters and split across cell types. (h) Stacked violin and Dot plot are the key visualizations of selected cell markers across cell types. They highlight their selective expression and validates the scRNAseq approach and visualization method. (i) Command Line Interface (CLI) exposed by mini Jupyter Notebook to provide maximal flexibility of doing various analytics on the whole or sliced single cell dataset.

1. Tang, F. et al. mRNA-Seq whole-transcriptome analysis of a single cell. Nature Methods 6, 377-382 (2009).
2. Svensson, V., da Veiga Beltrame, E. & Pachter, L. A curated database reveals trends in single-cell transcriptomics. bioRxiv, 742304 (2019).
3. Rodriques, S.G. et al. Slide-seq: A scalable technology for measuring genome-wide expression at high spatial resolution. Science 363, 1463 (2019).
4. Chan Zuckerberg Initiative chanzuckerberg/cellxgene: An interactive explorer for single-cell transcriptomics data. Available at https://chanzuckerberg.github.io/cellxgene/ ([Accessed: 5 May 2020]).
5. Cakir, B. et al. Comparison of visualization tools for single-cell RNAseq data. NAR Genomics and Bioinformatics 2 (2020).
6. Wolf, F.A., Angerer, P. & Theis, F.J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biology 19, 15 (2018).
7. Li, K. et al. figureComposer: A web-based interactive multi-panel bio-infographic designing tool. bioRxiv, 2020.2003.2004.976589 (2020).

Demo site: https://cellxgenevip-ms.bxgenomics.com

Pre-print: https://www.biorxiv.org/content/10.1101/2020.08.28.270652v1

Installation instruction

0. Install anaconda python 3.7 if not available on server

- Anaconda3-2020.02-Linux-x86_64.sh

1. create and enable conda environment

git clone https://github.com/interactivereport/cellxgene_VIP.git
cd cellxgene_VIP
conda env create -n <env name, such as: VIP> -f VIP.yml
conda activate <env name, such as: VIP>
or
source activate <env name>

2. Install cellxgene by run the config.sh in the folder

./config.sh

3. Run cellxgene by specifiying a h5ad file storing scRNA-seq data along with a host and a port, use "ps" to find used ports to spare, see https://chanzuckerberg.github.io/cellxgene/posts/launch for details.

ps -ef | grep cellxgene
Rscript -e 'reticulate::py_config()'
# Run the following command if the output of the above command is not the python in your env.
export RETICULATE_PYTHON=`which python`
cellxgene launch --host <xxx> --port <xxx> --disable-annotations --verbose <h5ad file>

4. From web browser (Chrome is preferred, Version 87.0.4280.88 or 87.0.4280.141 is used), access http(s)://host:port

You should be able to see this in Console of Chrome Developer Tools if everything is right.

note: while spinning up the cellxgene from HPC, do NOT use qlogin. ssh directly to the server.

Updating

./update.VIPInterface.sh # if "interface.html" or "VIPInterface.py" is modified, often.

./update.index_template.sh # if jsPanel is modified, very rare.

Note on installtion of packages

pandoc, install R Studio Server https://rstudio.com/products/rstudio/download-server/ and then,

ln -s /usr/lib/rstudio-server/bin/pandoc/pandoc /usr/bin

Install R and R Packages at system level (root or sudo privilege is needed)

R: https://cran.r-project.org/web/packages/arrow/index.html

In command line:

export LIBARROW_MINIMAL=false

Then in R:

install.packages("arrow")

R packages needed for volcano plot in DEG (Differentially Expressed Genes) analysis

> library(ggplot2)
> library(ggrepel)
> library(ggrastr)

Seurat ver3.0.2

mkdir /usr/lib64/R/library/Seurat_3.0.2/
$ R
>devtools::install_github(repo = 'satijalab/seurat@v3.0.2', dependencies=FALSE, force=TRUE, lib ='/usr/lib64/R/library/Seurat_3.0.2/')
>library(Seurat, lib.loc = '/usr/lib64/R/library/Seurat_3.0.2/')

Packages needed for CLI.

$ conda install ipykernel
$ pip install rpy2  # If you want to use globally installed R and packages
or
$ conda install rpy2  # R and R packages will be installed locally