Skip to contents

Installation

Source: vignettes/articles/manual-00-installation.Rmd
manual-00-installation.Rmd

Manual

Requirements

  • Operating system: Windows, macOS, or Linux.
  • R: the current release, from CRAN. Minimum is 4.1; on older releases some packages must be built from source, which is slower and requires additional tools.
  • R Editor (recommended): RStudio or Positron.
  • Quarto: bundled with recent RStudio and Positron releases; otherwise install it from quarto.org. Required only for rendering notebooks.

Installing and updating MRMhub

Important. Quit and reopen all RStudio/Positron sessions first. Many installation issues are caused by R packages that are loaded in another session while the installer tries to update them.

In a fresh R session:

# mrmhub from the MRMhub R-universe repository, dependencies from CRAN
install.packages("mrmhub",
                 repos = c("https://slinghub.r-universe.dev",
                           "https://cloud.r-project.org"))

The same command installs later updates.

Confirm the installation:

If R is older than the version the package was built with, the warning package 'mrmhub' was built under R version ... appears. It can be ignored: mrmhub contains no compiled code, so the version difference has no effect. If the package does not load, see Troubleshooting or the ZIP-file method below. For optional features, see Optional packages.

Alternative installation methods

If R-universe is unreachable, MRMhub can be installed from GitHub. Both routes build the package locally.

if (!requireNamespace("pak", quietly = TRUE)) install.packages("pak")
pak::pak("SLINGhub/MRMhub")

pak may fail because of a network or proxy restriction, or a request for build tools such as the Xcode command line tools on macOS. In that case, use remotes:

if (!requireNamespace("remotes", quietly = TRUE)) install.packages("remotes")
remotes::install_github("SLINGhub/MRMhub")

Installing from a downloaded repository ZIP file

If the methods above fail or R cannot reach the repositories, download a copy of the repository and install MRMhub from it:

  1. Open https://github.com/SLINGhub/MRMhub in a browser.
  2. Click the green Code button and choose Download ZIP (or download https://github.com/SLINGhub/MRMhub/archive/refs/heads/main.zip directly).
  3. Unzip the file. It expands to a folder named MRMhub-main.
  4. In a fresh R session, install from that folder with remotes:
remotes::install_local("path/to/MRMhub-main")

Replace path/to/MRMhub-main with the actual location of the unzipped folder, e.g. "~/Downloads/MRMhub-main".

Optional packages

To keep the base installation light, a number of specialised functions in mrmhub rely on additional packages that are not installed upfront. When such a function is called and its package is not yet installed, mrmhub offers to install it.

Function Optional package Enables
plot_runscatter() qpdf; mirai, carrier only for multi-threaded PDF output
plot_qc_summary_overall(with_venn = TRUE) ggvenn, patchwork Venn diagram of features excluded by QC criteria
plot_matrixeffects(), plot_interference_correction() ggbeeswarm beeswarm/quasirandom point layers
correct_isotopic_interferences(), calc_average_molweight() enviPat isotope-pattern and molecular-weight calculation
correct_batch_combat() sva ComBat batch correction
correct_batch_serrf() ranger SERRF batch correction
correct_drift_gam() mgcv GAM-based drift correction
build_workflow() shiny, bslib interactive workflow-builder app
save_dataset_summarizedexperiment() SummarizedExperiment, S4Vectors, lipidr (Bioconductor) export to a SummarizedExperiment
Lipid-name parsing (isotope correction, lipid plots) rgoslin (Bioconductor) parse and normalise lipid shorthand
get_response_curve_stats() lancer (GitHub) only for specific response-curve metrics

To install all of them upfront, run the following in a fresh R session (again, with all other RStudio/Positron sessions closed):

pak::pak(c(
  "sva", "ranger", "mgcv", "enviPat",
  "qpdf", "mirai", "carrier", "shiny", "bslib",
  "ggvenn", "patchwork", "ggbeeswarm",
  "rgoslin", "lipidr", "SummarizedExperiment", "S4Vectors",
  "SLINGhub/lancer"
))

Alternatively, the CRAN packages can be installed with install.packages(). For the Bioconductor ones (sva, rgoslin, lipidr, SummarizedExperiment, S4Vectors), use pak as shown above: it selects the Bioconductor release paired with the installed R version. BiocManager may instead resolve to an older release and report that the current one requires a newer R. That message is advisory, and upgrading R is not necessary.

Troubleshooting

If the installation seemed successful but the package does not load, or loads with errors, run check_setup() to report the R version and flag any missing dependencies:

mrmhub::check_setup()

Frequent installation errors

See Troubleshooting & FAQ for a detailed list of errors and resolutions.

Error Cause Fix
namespace 'rlang' is already loaded An old dependency is still loaded in the session Restart R (Ctrl+Shift+F10) and retry the install
pak unavailable or failing pak not installed, or its cache is stale Install with remotes::install_github("SLINGhub/MRMhub")
cannot open URL Firewall or proxy blocking GitHub options(download.file.method = "libcurl"), or install from a downloaded repository ZIP file
Could not resolve host: api.github.com Firewall or proxy blocking the GitHub API Set http_proxy/https_proxy, or install from a downloaded repository ZIP file
SSL certificate problem: self signed certificate in certificate chain Corporate proxy inspecting TLS traffic Use remotes (it uses the system certificate store), or install your organisation’s root certificate
package 'mrmhub' was built under R version … The R-universe binary is built under the current R patch release, newer than the R in use Cosmetic; mrmhub contains no compiled code. Update R to the current release to silence it
there is no package called 'mrmhub' Install did not finish Scroll up for the real error, then retry the install
ERROR: Rtools is required, or packages start compiling (Windows) A dependency has no binary for your R version Update R, or set options(install.packages.compile.from.source = "never"). If a source build is unavoidable, install Rtools matching your R version
Bioconductor version '3.23' requires R version '4.6' BiocManager selected a Bioconductor release newer than the installed R Install the Bioconductor packages with pak instead: it selects the release paired with your R. Upgrading R is not required
clang: error:, Could not find tools necessary to compile a package, or No developer tools were found (macOS) Xcode command line tools missing Run xcode-select --install in a Terminal, accept the dialog, retry in a fresh session, or install with remotes instead of pak
cannot find -lcurl (Linux) System libraries missing sudo apt install libcurl4-openssl-dev libxml2-dev libssl-dev libfontconfig1-dev

Next steps