This function calculates different metrics to evaluate the integration of scRNA expression matrices in a new dimension. Its a wrapper function around scib batch correction metrics
Usage
DO.EvalIntegration(
sce_object,
label_key = "annotation",
batch_key = "orig.ident",
type_ = "embed",
pcr_covariate = "orig.ident",
pcr_n_comps = 30,
scale = TRUE,
verbose = FALSE,
n_cores = 10,
assay = "RNA",
integration = "INTEGRATED.CCA",
kBET = TRUE,
cells.use = NULL,
subsample = NULL,
min_per_batch = NULL,
all_scores_silhouette = FALSE,
...
)Arguments
- sce_object
Seurat or SCE object.
- label_key
character, Annotation column
- batch_key
character, Sample column
- type_
character, default: "embed"
- pcr_covariate
character, covariate column for pcr
- pcr_n_comps
integer, number of components for pcr
- scale
boolean, default: TRUE
- verbose
boolean, defult: FALSE
- n_cores
integer, Number of cores used for calculations
- assay
character, Name of the assay the integration is saved in
- integration
character, Name of the integration to evaluate
- kBET
boolean, if kBET should be run
- cells.use
vector, named cells to use for kBET subsetting
- subsample
float, for starified subsampling,
- min_per_batch
integer, minimum number of cells per batch
- all_scores_silhouette
boolean, define if all scores of silhouette return
- ...
Additionally arguments for kBET
Examples
if (FALSE) { # \dontrun{
sce_data <-
readRDS(system.file("extdata", "sce_data.rds", package = "DOtools"))
DO.EvalIntegration(
sce_object = sce_data,
label_key = "annotation",
batch_key = "orig.ident",
type_ = "embed",
pcr_covariate = "orig.ident",
pcr_n_comps = 30,
scale = TRUE,
verbose = FALSE,
n_cores = 10,
assay = "RNA",
integration = "INTEGRATED.CCA",
kBET = TRUE,
cells.use = NULL,
subsample = NULL,
min_per_batch = NULL,
all_scores_silhouette = FALSE
)
} # }