impute_across_panels: Impute non-overlapping channels for whole data sets

In shdam/cyCombine: Robust integration of single-cell cytometry datasets within and across technologies

Impute non-overlapping channels for whole data sets

Description

This function imputes the expression values for a whole dataset based on the overlapping markers contained in another dataset. The purpose is to be able to merge multi-panel data or to merge differently run datasets that have non-overlapping markers

Usage

impute_across_panels(
  dataset1,
  dataset2,
  overlap_channels,
  impute_channels1,
  impute_channels2,
  xdim = 8,
  ydim = 8,
  seed = 482
)

Arguments

dataset1	Dataframe with expression values 1
dataset2	Dataframe with expression values 2
overlap_channels	Channels (markers) that overlap between impute_for and complete_obs, which can be used to base imputation on
impute_channels1	Channels to impute and add to the impute_for dataset1 (must be present in dataset2)
impute_channels2	Channels to impute and add to the impute_for dataset2 (must be present in dataset1)
xdim	The x-dimension size of the SOM.
ydim	The y-dimension size of the SOM.
seed	The seed to use when creating the SOM.

See Also

Other merging: salvage_problematic()

Examples

## Not run: 
dfs_imputed <- impute_across_panels(dataset1 = df_panel1, dataset2 = df_panel2,
 overlap_channels = intersect(df_panel1_markers, df_panel2_markers),
 impute_channels1 = unique_df_panel2_markers, , impute_channels2 = unique_df_panel1_markers)

## End(Not run)

shdam/cyCombine documentation built on June 12, 2022, 11:29 p.m.