scCTS: identifying the cell type specific marker genes from population-level single-cell RNA-seq

scCTS is an R package for the statistical modeling of the gene differential expression (DE) in scRNA-seq data. It identifies cell-type specific genes (markers) that consistently appear in historical population-level scRNA-seq (scRNA-seq) data. scCTS is built on top of the R package SingleCellExperiment and supports parallel computation.

Except from our proposed method, scCTS also provides a common interface for classic DE methods such as the Wilcoxon rank-sum test, t-test and DESeq2.

Installation

You can install scCTS from GitHub using the devtools package:

if (!require("devtools", quietly = TRUE))
    install.packages("devtools")
    
devtools::install_github('ToryDeng/scCTS', dependencies=T, build_vignettes = T)
library(scCTS)

Getting started

Here we give a simple example to demonstrate how to run scCTS. Once the package is installed, you can load the simulated dataset included in the package:

data(sim.sce)

sim.sce is a SingleCellExperiment object with 200 genes and 10,000 cells. Next, you can run scCTS with a single line of code:

res <- scCTS(sim.sce, subject.rep='subject', celltype.rep='celltype', numCores=2)

Some explanations about the parameters:

• subject.rep: The name of the column that stores subject labels of cells in the colData slot.
• celltype.rep: The name of the column that stores cell type labels in the colData slot.
• numCores: Number of cores for parallel computation.

In the tested environment, the code finishes running within 10 seconds. The return value res is a list containing lists for each cell type. Each list contains posterior probabilities of genes and parameter estimations for a particular cell type. For example, you can extract the posterior probabilities of genes to show DE in celltype1 using the following code:

res$celltype1$pp.d1

For more details about how to run scCTS and classic DE methods, please refer to vignette("scCTS").

Reproducibility

If you want to reproduce results shown in the paper, please refer to the directory reproducibility/ in this repo.

Tested environments

Environment 1

• CPU: AMD Ryzen Threadripper 3990X 64-Core Processor
• Memory: 256GB
• System: Ubuntu 20.04.6 LTS
• R version: 4.3.0

Environment 2

• CPU: Intel(R) Xeon(R) Gold 6240R CPU @ 2.40GHz
• Memory: 256GB
• System: Ubuntu 22.04.3 LTS
• R version: 4.4.0

Citation

Coming soon.