Find region neighbors

This function allows you to automatically identify neighbors of a selected region.

Usage

RegionNeighbors(object, ...)

# S3 method for default
RegionNeighbors(
  object,
  spots,
  mode = c("outer", "inner", "inner_outer", "all_inner_outer"),
  verbose = FALSE,
  ...
)

# S3 method for Seurat
RegionNeighbors(
  object,
  column_name,
  column_labels = NULL,
  mode = c("outer", "inner", "inner_outer", "all_inner_outer"),
  column_key = NULL,
  verbose = TRUE,
  ...
)

Arguments

object

An object

...

Arguments passed to other methods

spots

A character vector with spot IDs present in `spatnet`

mode

Select mode (see details)

verbose

Print messages

column_name

string specifying a column name in your meta data with labels, e.g. clusters or manual selections

column_labels

character vector with labels to find nearest neighbors for. These labels need to be present in the meta data columns specified by column_name

column_key

prefix to columns returned in the Seurat object

Value

An object with labeled region neighbors

Seurat

If a Seurat object is provided, the RegionNeighbors takes a meta data column (chosen with column_name) with categorical labels, finds the nearest neighbors of spots for a selected group in this columns (chosen with column_labels) and returns new meta data column with labels for the nearest neighbors of the selected group. If no column_labels are specified, the method will return a column for each separate category in the column_name vector.

Note that the prefix to the returned column names will be selected based on the mode. You can overwrite this behavior by manually setting column_key.

Below is some additional information about the behavior of different modes:

• return outer border (default): mode="outer"

• return inner border: mode="inner"

• return inner and outer borders: mode="inner_outer"

• return all selected spots and outer border: mode="all_inner_outer"

default method

The default method takes a list of spatial networks generated with GetSpatialNetwork together with a vector of spot IDs and returns the spot IDs for border spots. The behavior for border spot selection is determined by the mode.

See also

Other spatial-methods: CorSpatialFeatures(), CutSpatialNetwork(), DisconnectRegions(), GetSpatialNetwork(), RadialDistance(), RunLabelAssortativityTest(), RunLocalG(), RunNeighborhoodEnrichmentTest()

Author

Ludvig Larsson

Examples

library(semla)
library(dplyr)

se_mbrain <-
  readRDS(system.file("extdata",
  "/mousebrain/se_mbrain",
  package = "semla"))

# Create Seurat object
se_mbrain <- se_mbrain |>
  ScaleData(verbose = FALSE) |>
  RunPCA(verbose = FALSE) |>
  FindNeighbors(verbose = FALSE) |>
  FindClusters(verbose = FALSE)

# Find neighbors to cluster 10
se_mbrain <- RegionNeighbors(se_mbrain,
                             column_name = "seurat_clusters",
                             column_labels = "10")
#> ℹ Finding neighboring spots for '10'
#> →   Excluding neighbors from the same group
#> →   74 neighbors left
#> →   Returning neighbors

# Plot cluster 10 and its neighbors
se_mbrain$selected_clusters <- se_mbrain[[]] |>
  mutate(across(where(is.factor), as.character)) |>
  mutate(cl = case_when(seurat_clusters %in% "10" ~ seurat_clusters,
                        TRUE ~ NA_character_)) |>
  pull(cl)

MapLabels(se_mbrain, column_name = "selected_clusters") |
  MapLabels(se_mbrain, column_name = "nb_to_10")


# \donttest{
# Find neighbors to clusters 8 and 10
se_mbrain$selected_clusters <- se_mbrain[[]] |>
  mutate(across(where(is.factor), as.character)) |>
  mutate(cl = case_when(seurat_clusters %in% c("8", "10") ~ seurat_clusters,
                        TRUE ~ NA_character_)) |>
  pull(cl)
se_mbrain <- RegionNeighbors(se_mbrain,
                             column_name = "seurat_clusters",
                             column_labels = c("8", "10"))
#> ℹ Finding neighboring spots for '8'
#> →   Excluding neighbors from the same group
#> →   112 neighbors left
#> →   Returning neighbors
#> ℹ Finding neighboring spots for '10'
#> →   Excluding neighbors from the same group
#> →   74 neighbors left
#> →   Returning neighbors

# Plot cluster 8, 10 and its neighbors
library(patchwork)
MapLabels(se_mbrain, column_name = "selected_clusters") +
  MapLabels(se_mbrain, column_name = "nb_to_8") +
  MapLabels(se_mbrain, column_name = "nb_to_10") +
  plot_layout(design = c(area(1, 1, 1, 1),
                         area(1, 2, 1, 2),
                         area(1, 3, 1, 3)))


# it is also possible to pass additional parameters to GetSpatialNetwork
# to make it find more neighbors at a larger distances
se_mbrain <- RegionNeighbors(se_mbrain,
                             column_name = "seurat_clusters",
                             column_labels = "10",
                             nNeighbors = 40,
                             maxDist = Inf)
#> ℹ Finding neighboring spots for '10'
#> →   Excluding neighbors from the same group
#> →   860 neighbors left
#> →   Returning neighbors
MapLabels(se_mbrain, column_name = "nb_to_10")

# }