nbahti/popsicleR

popsicleR, a flexible R package aimed at providing most of the necessary quality controls and preliminary information for complete and detailed analysis of single cell RNA-seq data.

popsicleR

popsicleR is a R package that combines methods implemented in widely used pipelines to interactively perform all major pre-processing and QC steps of scRNA-seq data analysis. The package is composed of seven functions capable of performing exploration of quality-control metrics, filtering of low-quality cells, data normalization, removal of technical and biological biases, and some basic analysis as detection of differentially expressed genes, cell clustering and cell annotation. During each step of the analysis, popsicleR interactively guides the user with colored text messages and saves in dedicated folders a variety of plots to investigate several QC metrics and assess the impact of filtering and regression parameters on the identification and classification of cell populations.

Key features of popsicleR include:

1. Use as input of files from either the Cell Ranger pipeline of 10X Genomics or a feature-barcode matrix of raw counts generated from any microfluidic-, microwell plates-, or droplet-based scRNA-seq technology
2. Output of graphs and colored text messages to interactively guide users along each step of the analysis
3. Inclusion of common single-cell visualisations (as density, scatter, and violin plots and low-dimensionality embeddings) to investigate QC metrics and pre-processing parameters
4. Export of visualisations as PDF images for presentation or publication use

Contact:

silvio.bicciato@unimore.it; mattia.forcato@unimore.it

Citation:

F. Grandi, J. Caroli, O. Romano, M. Marchionni, M. Forcato, S. Bicciato, popsicleR: a R Package for pre-processing and quality control analysis of single cell RNA-seq data, Journal of Molecular Biology (2022), doi: 10.1016/j.jmb.2022.167560

Table of Contents

• System requirements
• Installation in R
• Installation through conda
• Tutorial on example data

System requirements

• R version: >= 4.0.0
• Dependencies: ape, celldex, clustree, corrplot, crayon, dplyr, future, ggExtra, ggplot2, ggplotify, gtools, grid, gridExtra, limma, magrittr, patchwork, pheatmap, neldermead, RANN, RColorBrewer, reticulate, R.utils, scDblFinder, scMCA, session, shinythemes, umap, Seurat, and SingleR.

Installation in R

Before installing popsicleR, users must run the following codes to install packages required as dependencies from CRAN, CRAN archive, Bioconductor, and Github. During this stage, we recommend skipping updates and the installation of Rtools when prompted by R.

CRANdep <- c("Seurat","reticulate","R.utils","dplyr","ggplot2","clustree","ape","gtools",
"future","grid","gridExtra","magrittr","limma","patchwork",
"crayon","ggExtra","RColorBrewer","ggplotify","RANN","umap",
"celldex","curl","httr","lattice","shinythemes","usethis","rcmdcheck",
"roxygen2","rversions","devtools","pheatmap","BiocManager","corrplot")
newPackages <- CRANdep[!(CRANdep %in% installed.packages()[,"Package"])]
if(length(newPackages)){install.packages(newPackages)}

CRANarcdep <- c("Matrix","optimbase","optimsimplex","neldermead","session")
newPackages <- CRANarcdep[!(CRANarcdep %in% installed.packages()[,"Package"])]
if(length(newPackages)){
	packagesurl <- c("https://cran.r-project.org/src/contrib/Archive/Matrix/Matrix_1.3-2.tar.gz",
	"https://cran.r-project.org/src/contrib/Archive/optimbase/optimbase_1.0-9.tar.gz",
	"https://cran.r-project.org/src/contrib/Archive/optimsimplex/optimsimplex_1.0-7.tar.gz",
	"https://cran.r-project.org/src/contrib/Archive/neldermead/neldermead_1.0-11.tar.gz",
	"https://cran.r-project.org/src/contrib/Archive/session/session_1.0.3.tar.gz")
	for (i in 1:length(newPackages)){
		source_repo <- packagesurl[grep(newPackages[i], packagesurl)]
		install.packages(source_repo, repos=NULL, type="source")
	}
}

BioCdep <- c("SingleR","limma","BiocFileCache","AnnotationHub",
"ExperimentHub","celldex","scDblFinder")
newPackages <- BioCdep[!(BioCdep %in% installed.packages()[,"Package"])]
if(length(newPackages)){BiocManager::install(newPackages)}

if(!"scMCA"%in% installed.packages()[,"Package"]){devtools::install_github("ggjlab/scMCA")}

Once installed all dependencies, popsicleR can be installed with the following script:

devtools::install_github("bicciatolab/popsicleR")

In case of any issue with installation of popsicleR via install_github, it is possible to download the package.tar.gz using the bash command:

 wget https://github.com/bicciatolab/popsicleR/archive/main.tar.gz

Finally, extract the main directory (if necessary, rename the package folder from "popsicleR-main" to "popsicleR") and install popsicleR from the local repository with the following script:

install.packages("/path/to/package_directory", repos = NULL, type="source")

Installation through conda

If Anaconda is already installed, a virtual environment for popsicleR can be set either manually, installing all packages one by one, or automatically adding only selected packages once created the environment.

The following comands allow setting the popsicleR environment.

Create a popsicleR environment and install all anaconda available packages automatically

To create the popsicleR environment on a Linux machine, open the terminal and run:

conda create -n popsicleR -c conda-forge r-base=4.0.3 r-umap=0.2.7.0 r-neldermead=1.0_11 r-rann=2.6.1 r-rcolorbrewer=1.1_2 r-ggextra=0.9 r-ggplotify=0.1.0 r-crayon=1.4.0 r-patchwork=1.1.1 r-magrittr=1.5 r-gridextra=2.3 r-dplyr=1.0.4 r-ggplot2=3.3.3 r-devtools=2.3.2 r-r.utils=2.10.1 r-future=1.21.0 r-reticulate=1.18 r-pheatmap=1.0.12 r-shinythemes=1.2.0 r-rcurl=1.98_1.2 r-corrplot=0.92 r-locfit=1.5_9.4 r-clustree=0.4.4 r-ape=5.6 r-seuratobject=4.0.4 r-sessioninfo=1.1.1 r-seurat

Install environment packages

From command line, use the following conda commands to install packages from other channels:

conda install -n popsicleR -c r r-magrittr
conda install -n popsicleR -c bioconda bioconductor-limma=3.46.0

Since not all required packages are provided in anaconda.org, some packages must be intalled directly from the R console, as described in install packages in R.

Install popsicleR environment through a .yml file

conda popsicleR environment can also be extracted from a popsicleR.yml file. In this case, all anaconda required packages will be automatically installed. After downloading the popsicleR.yml file in the working directory (or on a specific file_path), run:

conda env create -n popsicleR -f popsicleR.yml

Since not all required packages are provided in anaconda.org, some packages must be intalled directly from the R console, as described in install packages in R.

Install packages in R

Once created the environment, access it through the command:

conda activate popsicleR

and install session, SingleR, celldex, scDblFinder and scMCA packages using:

install.packages("session")

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("SingleR")
BiocManager::install("celldex")
BiocManager::install("scDblFinder")

devtools::install_github("ggjlab/scMCA")

Finally, use the following scripts to install popsicleR from Github :

devtools::install_github("bicciatolab/popsicleR")

or from a local repository:

install.packages("/path/to/package_directory", repos = NULL, type="source")