# recommended
BiocManager::install("gdagstn/cellula")
remotes::install_github("gdagstn/cellula")
devtools::install_github("gdagstn/cellula")Installing cellula
Installation
Use remotes, devtools or BiocManager to install. The BiocManager installation is probably the best one, as it will be able to fetch all the Bioconductor dependencies automatically.
Dependencies
cellula is dependency-heavy, which is not something I’m proud of, but makes sense considering this is a wrapper to a series of different analytical approaches. However, to keep the actual dependency tree lighter for the package itself, as many dependencies as possible have been moved to the Suggests part of the DESCRIPTION.
This way, necessary packages for specific methods will be checked only when executing functions that depend on those packages. This means that cellula won’t ask you to install Seurat until the moment you want to use Seurat integration or module scoring. This may be a little more tedious than having all the dependencies ready at the beginning, but it feels like a better solution than shoehorning 30+ dependencies at first install. Moreover, if you’ve been working with single cell analysis for a bit, you probably have most of these already installed.
Should you want to install all optional dependencies at once before installing cellula, you can install them as follows:
BiocManager::install(c("DropletUtils", "scDblFinder", "AUCell", "harmony", "GSVA",
"Seurat", "rliger", "tricycle", "UCell", "slingshot",
"cole-trapnell-lab/monocle3", "destiny", "carmonalab/STACAS",
"welch-lab/RcppPlanc"))Moreover, if you want to install Bioconductor non-optional dependencies separately (e.g. to install cellula using remotes) you can install them as follows:
BiocManager::install(c("SingleCellExperiment", "SummarizedExperiment",
"BiocNeighbors", "scran", "scater", "scuttle", "bluster",
"batchelor", "BiocParallel", "S4Vectors", "TSCAN"))