An R package for simple transcriptome meta-analysis for identifying stress-responsive genes
Stress Response score (SRscore) is a stress responsiveness measure for transcriptome datasets and is based on the vote-counting method. The SRscore is determined to evaluate and scores genes on the basis of the consistency of the direction of their regulation (Up-regulation, Down-regulation, or No changed) under stress conditions across the analyzed, multiple research projects. This package is based on the HN-score of Tamura and Bono (2022), and can calculate both the original method and the calculation method we have extended (Fukuda et al. 2025).
SRscore package requires not only R but also Rtools to be installed beforehand. Please download Rtools from https://cran.r-project.org/bin/windows/Rtools/ and install it before installing SRscore package.
On recent macOS versions, XQuartz (X11) is not installed by default.
Missing XQuartz may cause errors such as:
cannot load shared object 'R_X11.so'
To avoid this, install XQuartz from https://www.xquartz.org/ . After installation, restart your system or log out/in, then proceed to "One-Stop" installation.
There are several system-level software packages required to install the R packages that SRscore package depends on. Before installing SRscore package, please install these prerequisites using the following command in your bash shell.
sudo apt install -y libcurl4-openssl-dev libssl-dev libxml2-dev libfreetype6-dev libpng-dev libtiff5-dev libjpeg-dev libfontconfig1-dev libharfbuzz-dev libfribidi-dev libwebp-dev pkg-config build-essential gfortran libcairo2-dev libglpk-dev pandoc libmagick++-dev
Pandoc is required if you install SRscore package from this GitHub repository rather than CRAN, and build the documentation locally.
In that case, please follow Instruction https://pandoc.org/installing.html and install Pandoc for your operating system.
"Installing from CRAN" is strongly recommended, particularly on Windows and macOS.
SRscore package is available on CRAN https://cran.r-project.org/web/packages/SRscore/ . To install it from CRAN, please run the following command in your R console.
# When running Rscript install.R from the command line, it fails if no mirror is set. Use this line to the top of the script would solve it.
options(repos = c(CRAN = "https://cloud.r-project.org/"))
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
# Install Bioconductor dependencies first
BiocManager::install(c("ComplexHeatmap", "clusterProfiler", "org.At.tair.db", "genefilter", "BiocStyle"))
install.packages("SRscore")
# When running Rscript install.R from the command line, it fails if no mirror is set. Use this line to the top of the script would solve it.
options(repos = c(CRAN = "https://cloud.r-project.org/"))
# devtools has a very large dependency tree that often breaks in Conda environments. Recommending remotes::install_github() is much safer and lighter for most users.
if (!requireNamespace("remotes", quietly = TRUE))
install.packages("remotes")
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
# Install Bioconductor dependencies first
BiocManager::install(c("ComplexHeatmap", "clusterProfiler", "org.At.tair.db", "genefilter", "BiocStyle"))
# Install SRscore (set build_vignettes = FALSE if Pandoc is not available)
pandoc_path <- Sys.which("pandoc")
if (pandoc_path != "") {
install.packages(c("magick", "DT"))
remotes::install_github("fusk-kpu/SRscore", build_vignettes = TRUE)
} else {
remotes::install_github("fusk-kpu/SRscore", build_vignettes = FALSE)
}
NOTE: option "build_vignettes = FALSE" is highly recommended.
In some computing environments—particularly headless Linux systems, Conda-based setups, or servers with limited user permissions—installation of optional packages used in example workflows (e.g., enrichment analysis with clusterProfiler) may fail.
Please note that these packages are not required for the core functionality of SRscore. The SRscore package itself can be installed and used independently of these optional dependencies.
Installation failures of optional packages may occur due to:
Missing system-level libraries (e.g., BLAS/OpenBLAS, graphics or font libraries)
Restricted write permissions to system library paths
Differences between CRAN and Bioconductor dependency versions
Minimal or containerized Linux environments
If such issues arise, we recommend the following:
Install SRscore from CRAN using the standard command:
install.packages("SRscore")
Skip building vignettes or optional examples:
remotes::install_github("fusk-kpu/SRscore", build_vignettes = FALSE)
Install optional Bioconductor packages only if needed:
install.packages("BiocManager")
BiocManager::install("clusterProfiler")
In restricted environments, it may also be necessary to install additional system libraries or use a user-level R library path.
library(SRscore)
browseVignettes("SRscore")
The SRscore package is designed to facilitate meta-analysis
methods based on vote-counting.
It contains three main functions for calculating the SRscore,
which represents a numerical value indicating a gene's stress
responsiveness among multiple studies.
Using the expand_by_groups() function,
it is possible to generate a table pairing all possible combinations of two groups,
which can be arranged in two columns. To mitigate batch effects,
the function only generates pairs among samples within a given dataset (e.g., NCBI GEO series).
When the table thus acquired is used as an input to execute the calc_SRratio() function,
this function calculates a value designated the Stress Response ratio (SRratio) and,
which is stored in an SRratio matrix (gene × sample).
SRratio represents the gene expression level and is calculated similarly to a log2 fold change.
Using this matrix as an input, executing the calc_SRscore() function yields a gene-specific SRscore.
The primary feature of the SRscore package is its capacity to perform cross-comparative analysis of multiple datasets and to estimate consistent changes in gene expression levels. Commencing with the import of metadata and expression data, the package implements a sequential workflow that includes inter-group comparisons within each dataset, calculation of integrated scores via meta-analysis, and visualization and export of the results.
MIT