Installing PRIME

This guide describes how to install the PRIME R package and its dependencies.

PRIME depends on a mix of CRAN, Bioconductor, and GitHub packages.

Quick start (recommended)

Start an R session:

1) Install BiocManager

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

2) Install Bioconductor dependencies (recommended: start with CAGEfightR)

PRIME imports CAGEfightR. Installing it will pull in most Bioconductor dependencies automatically.

BiocManager::install("CAGEfightR", ask = FALSE, update = FALSE)

3) If installation fails due to `pryr` (archived on CRAN)

Some dependency stacks require pryr (>= 0.1.3). pryr was archived on CRAN on 2026-01-30.

If you see an error mentioning pryr, install it from the CRAN archive:

if (!requireNamespace("remotes", quietly = TRUE)) {
  install.packages("remotes")
}

remotes::install_url(
  "https://cran.r-project.org/src/contrib/Archive/pryr/pryr_0.1.6.tar.gz",
  upgrade = "never"
)

Then retry installing CAGEfightR:

BiocManager::install("CAGEfightR", ask = FALSE, update = FALSE)

4) Install devtools (or remotes)

You can install PRIME using either devtools or remotes. If you already installed remotes above, you can skip installing devtools.

if (!requireNamespace("devtools", quietly = TRUE)) {
  install.packages("devtools")
}

5) Install `bcp` from GitHub

bcp is installed from GitHub:

devtools::install_github("swang87/bcp")

6) Install PRIME from GitHub

devtools::install_github("anderssonlab/PRIME")

Optional: install a small set of commonly-used CRAN packages

Most CRAN dependencies for PRIME will be installed automatically when you install PRIME. If you want to proactively install a few common packages (helpful on some systems), you can do:

install.packages(c(
  "assertthat",
  "Matrix",
  "data.table",
  "caTools",
  "dplyr",
  "foreach",
  "ggplot2",
  "igraph",
  "magrittr",
  "purrr",
  "stringr",
  "tibble",
  "tidyr",
  "tidyselect",
  "zoo"
))

Notes on R / Bioconductor versions

PRIME should be installable on newer versions of R. In general, you should use the Bioconductor release that matches your R version (BiocManager will normally pick the correct one automatically).
The PRIME GitHub Actions workflow used to build the pkgdown site is pinned to:
- R 4.2.2
- Bioconductor 3.16

If you need to reproduce that exact environment (e.g., for server installs or CI parity), you can explicitly pin Bioconductor like this:

BiocManager::install(version = "3.16", ask = FALSE, update = FALSE)

If you are using a newer R version, do not force Bioconductor 3.16; instead, install the Bioconductor version recommended for your R version.

Troubleshooting

System dependencies (Linux)

If you see compilation failures when installing packages from source, you may need system libraries (e.g. libcurl, openssl, xml2, etc.). The exact package names differ by Linux distribution.