This function will run the scVI Integration from the scVI python package. It includes all parameters from the actual python package and runs it by using an internal python script. The usage of a gpu is incorporated and highly recommended.
Usage
DO.scVI(
sce_object,
batch_key,
layer_counts = "counts",
layer_logcounts = "logcounts",
categorical_covariates = NULL,
continuos_covariates = NULL,
n_hidden = 128,
n_latent = 30,
n_layers = 3,
dispersion = "gene-batch",
gene_likelihood = "zinb",
get_model = FALSE
)Arguments
- sce_object
Seurat or SCE object with annotation in meta.data
- batch_key
meta data column with batch information.
- layer_counts
layer with counts. Raw counts are required.
- layer_logcounts
layer with log-counts. Log-counts required for calculation of HVG.
- categorical_covariates
meta data column names with categorical covariates for scVI inference.
- continuos_covariates
meta data column names with continuous covariates for scVI inference.
number of hidden layers.
- n_latent
dimensions of the latent space.
- n_layers
number of layers.
- dispersion
dispersion mode for scVI.
- gene_likelihood
gene likelihood.
- get_model
return the trained model.
Examples
sce_data <- readRDS(system.file("extdata", "sce_data.rds", package = "DOtools"))
# Run scVI using the 'orig.ident' column as the batch key
sce_data <- DO.scVI(sce_data, batch_key = "orig.ident")
#> ℹ Using the 'counts' assay as the X matrix