Skip to contents

sciCSR (Single-cell Inference of Class-Switch Recombination)

logo

Introduction

sciCSR is a R package designed to analyse class-switch recombination (CSR) in B cell single-cell RNA sequencing (scRNA-seq) data. CSR is a major process in B cell maturation whereby B cell changes the constant region (“isotype”) of their B cell receptor (BCR) in order to adapt their function to different tissue and biological contexts. Along with somatic hypermutation (SHM) where mutations are accumulated in the variable regions of BCR, CSR is a hallmark of the transitions of B cells from a naive state to acquire memory against antigens. In sciCSR we provides routines to extract information on CSR and SHM from scRNA-seq data and, if available, scBCR-seq (i.e. single-cell BCR repertoire profiling) data.

In theory CSR and SHM information can be used as alternatives to the popular RNA velocity method to infer transitions between B cells. sciCSR takes forward the extracted CSR/SHM information and inputs them to the CellRank method to infer transitions between B cells; this can be used directly to analyse patterns of B cell maturation in the data, by considering evidence from CSR and SHM, both of which are native to B cell biology, addressing the limitations of RNA velocity in analysing scRNA-seq datasets of mature cell types such as B cells found in circulation or in secondary lymphoid organs.

Installation

Note: the installation and use of the sciCSR package has been tested in Linux, MacOS and Windows machines. That said, there may still be uncatched issues in installation and/or usage - please log an issue in the github and we will look into them.

# install dependencies and sciCSR itself (requires 'devtools')
if (!require('devtools')){
  install.packages('devtools')
}
devtools::install_github(
  c("immunogenomics/harmony", "cellgeni/sceasy", "Fraternalilab/sciCSR")
)

The installation of the sciCSR package itself takes around 30 seconds on a modern standard Windows laptop.

sciCSR requires a few python packages to be installed (scanpy, scvelo, cellrank etc.) as its functionalities depend on these python packages. We recommend setting up a conda environment with these dependencies so that sciCSR exclusively calls packages from this environment; this will avoid problems with messing up your local installation of python (if any) if you have previously set up these packages for other uses. You don’t need to externally set this up - before the first time you run any examples/analyses with sciCSR, just run the following lines in R after finishing the sciCSR installation:

library(sciCSR)
prepare_sciCSR()

This will set up an environment named ‘scicsr’ with all the desired dependencies ready for your analysis. The code should detect any existing Anaconda/Miniconda installation and uses that to set up the environment; if not, it will install Miniconda at the R default location and set up the conda environment. This function takes around 5 minutes 30 seconds to set up all python dependencies on a modern standard Windows laptop.

Examples

Please consult the following vignettes: