Function to pass Default Arguments for each method
Usage
liana_defaults(
assay = "RNA",
assay.type = "logcounts",
expr_prop = 0.1,
seed = 1004,
complex_policy = "mean0",
parallelize = FALSE,
workers = 8,
permutation.params = NULL,
liana_pipe.params = NULL,
liana_call.params = NULL,
cellphonedb.params = NULL,
natmi.params = NULL,
sca.params = NULL,
connectome.params = NULL,
cytotalk.params = NULL,
logfc.params = NULL,
cellchat.params = NULL,
squidpy.params = NULL,
call_sca.params = NULL,
call_natmi.params = NULL,
call_connectome.params = NULL,
call_italk.params = NULL,
...
)
Arguments
- assay
Assay name passed to `call_italk`, `call_sca`, `call_cellchat`, and `call_connectome`
- assay.type
- the type of data to be used to calculate the means (logcounts by default), available options are: "counts" and "logcounts"
- expr_prop
minimum proportion of gene expression per cell type (0.1 by default). Note that when working with complexes, the minimum subunit proportion will be used for filtering.
- seed
random seed integer
- complex_policy
policy how to account for the presence of complexes.
- parallelize
whether to parallelize cellphonedb-like
- workers
number of workers to be called
- permutation.params
list of parameters passed to permutation methods
- liana_pipe.params
list of Parameters passed to NATMI
liana_pipe
- liana_call.params
list of Parameters passed to NATMI
liana_call
- cellphonedb.params
list of Parameters passed to liana's internal cellphonedb implementation
cellphonedb_score
- natmi.params
list of Parameters passed to liana's internal edge_specificity implementation
natmi_score
- sca.params
list of Parameters passed to liana's internal LRScore implementation
sca_score
- connectome.params
list of Parameters passed to liana's internal connectome's weight_sc implementation
connectome_score
- cytotalk.params
list of Parameters passed to liana's internal crosstalk scores implementation
cytotalk_score
- logfc.params
list of Parameters passed to liana's internal logFC implementation
logfc_score
- cellchat.params
list of Parameters passed to CellChat
call_cellchat
- squidpy.params
list of Parameters passed to Squidpy
call_squidpy
- call_sca.params
list of Parameters passed to SingleCellSignalR
call_sca
- call_natmi.params
list of Parameters passed to NATMI
call_natmi
- call_connectome.params
list of Parameters passed to Connectome
call_connectome
- call_italk.params
list of Parameters passed to iTALK
call_italk
Details
The default parameters for each method can also be overwritten by
manually passing a list of parameters for the appropraite method
liana_wrap
Further, each `get_*` method will by default obtain the default params passed
via liana_pipe
and liana_call
. This is done so that most steps
required for the calculation of these methods are undertaken only once.
NB! LIANA's internal methods are made consistent. There is no reason to pass specific parameters to any of them. Thus, it is best that one sticks to the non-nested parameters of this function (i.e. excluding `.params`), unless a very specific reason requires any of LIANA's internal parameters to be changed.
Examples
liana_path <- system.file(package = "liana")
# load testdata
testdata <- readRDS(file.path(liana_path , "testdata", "input", "testdata.rds"))
# get a `named` list with all default parameters passed to liana.
def_params <- liana_defaults()
# any of these can then be overwritten and passed to `...` in `liana_wrap`
# with the `.params` suffix to the parameter name type. For example,
liana_res <- liana_wrap(testdata,
permutation.params = list(nperms=2),
liana_pipe.params = list(test.type='wilcox'))
#> Expression from the `RNA` assay will be used
#> Running LIANA with `seurat_annotations` as labels!
#> Warning: 3465 genes and/or 0 cells were removed as they had no counts!
#> LIANA: LR summary stats calculated!
#> Now Running: Natmi
#> Now Running: Connectome
#> Now Running: Logfc
#> Now Running: Sca
#> Now Running: Cellphonedb