#Installation from source

NOTE: For Sanger farm22 users see instructions below

This will install the latest development version. These docs are for tglow-r v0.1.23 or later. The only tagged release (v0.1.2) is older than this, so for stability you can install a specific commit with the ref argument of remotes::install_git(), for example ref = "<commit sha>".

If you need to build some dependencies from source, make sure there is a BLAS/LAPACK, nlopt (nlopt), libxml2 (igraph) lib available if it isn't already, otherwise dependencies likely will not install. However this will depend heavily on your setup. Below is a minimal example using conda.

conda install -c conda-forge blas lapack nlopt libxml2
conda install -c conda-forge r-base r-arrow

arrow is needed to read the parquet files of tglow-pipeline v0.2.0. On Linux, installing it from CRAN can mean building it from source, so installing r-arrow with conda (or install.packages("arrow") before tglow-r) is usually easier.

Then launch R

if (!requireNamespace("remotes", quietly = TRUE)) {
  install.packages("remotes")
}
remotes::install_git("https://github.com/TrynkaLab/tglow-r.git")

This unlocks all of the core functionality

#Installing suggested packages

There are several suggested packages that enable additional functionality relating to visualizing images. These are outlined below.

#ggrastr - Rasterizing large plots

To enable rasterization of plots with many points (this is fully optional, and plotting will work without it)

# Optional if installing ggrastr to enable rasterization of plots with many points
conda install -c conda-forge r-ragg

Then launch R

install.packages("ggrastr")

#EBImage and ggiraph - Visualizing cells and Interactive plots

To enable visualizing images and making interactive plots EBImage, hdf5r, png, base64enc and ggiraph packages are required

install.packages("png")
install.packages("base64enc")
install.packages("ggiraph")
install.packages("hdf5r")

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

For detailed install instructions for ggiraph see: https://davidgohel.github.io/ggiraph/

#Installing on Sanger farm22 - latest dev version

A version compatible with the tglow-r softpack module comes pre-installed in /software/teamtrynka/installs/tglow-rlibs and can be loaded as such.

#Option 1: Using pre-installed version (recommended)

NOTE: The version here changes often at the moment, so might not be the most stable.

module load HGI/softpack/groups/cell_activation_tc/tglow-r/17

Then launch R.

library(tglowr, lib="/software/teamtrynka/installs/tglow-rlibs")

#Option 2: Installing into your personal library from git

Alternatively you can install if using R from the headnode or jammy64 directly from GitHub. This module should have all dependencies pre-installed.

module load HGI/softpack/groups/cell_activation_tc/tglow-r/17

Then launch R.

install.packages("remotes")
remotes::install_git("https://github.com/TrynkaLab/tglow-r.git")