Assign a cluster assignment to a muscadet object

Add the user selected cluster assignments to cells in a muscadet object. This function allows the user to choose the cluster assignments they consider to fit the data and their requirements, or cluster assignments based on other data and methods.

Usage

assignClusters(
  x,
  partition = NULL,
  clusters = NULL,
  mapping = NULL,
  redo_imputation = TRUE,
  knn_imp = 10
)

Arguments

x: A muscadet object (muscadet).
partition: Value specifying the clustering partition to choose from the muscadet object (numeric or character). It should be either the resolution or the k number of cluster (k) used for clustering depending on the clustering method (res_range or k_range with clusterMuscadet()). Should be provided if clusters is NULL.
clusters: A custom named vector of cluster assignments (vector). The vector names must match cell names in the muscadet object x, at least cluster assignments for all common cells must be provided if redo_imputation is set to true, otherwise, all cells within the muscadet object x must be provided. Should be provided if partition is NULL.
mapping: Optional named vector specifying how to remap cluster values (vector). The names of the vector correspond to the original cluster values, and the values are the remapped cluster values. For example, c("1" = 1, "2" = 1, "3" = 2, "4" = 3) would merge clusters 1 and 2 into 1, cluster 3 into 2, and cluster 4 into 3.
redo_imputation: Logical. If TRUE (default), reruns the imputation process to assign clusters to cells with missing data. This ensures that imputed clusters are updated if the clustering has changed due to remapping or to the use of custom clusters.
knn_imp: Integer specifying the number of nearest neighbors cells to use for imputation (integer). Must be a positive integer. Default is 10.

Value

A muscadet object updated with the user chosen cluster assignments in muscadet_obj$cnacalling$clusters.

Details

The clusters can be taken directly from the muscadet object clustering results with setting the parition argument (e.g. muscadet_obj$clustering$clusters[["0.8"]] for res=0.8).
A custom vector of cluster assignments can be attributed using the clusters argument.
Either way, the clusters assignments can be rearranged using the mapping argument.

Examples

if (FALSE) { # \dontrun{
# Load example muscadet object
data(muscadet_obj)

# Select clustering result for partition = 0.6
muscadet_obj <- assignClusters(muscadet_obj, partition = 0.6)
table(muscadet_obj$cnacalling$clusters)

# Assign custom clusters
set.seed(42)
cell_names <- Reduce(union, SeuratObject::Cells(muscadet_obj))
n1 <- sample(1:length(cell_names), 1)
n2 <- length(cell_names) - n1
custom_clusters <- setNames(c(rep.int(1, n1), rep.int(2, n2)), cell_names)
table(custom_clusters)
muscadet_obj <- assignClusters(muscadet_obj, clusters = custom_clusters)
table(muscadet_obj$cnacalling$clusters)

# Assign clusters with remapping
# example to remap from partition=0.8 with merging of clusters 2 and 3
clusters <- muscadet_obj$clustering$clusters[["0.8"]]
table(clusters) # 3 clusters
mapping <- c("1" = 1, "2" = 2, "3" = 2) # remap to 2 clusters

muscadet_obj <- assignClusters(muscadet_obj, clusters = clusters, mapping = mapping)
table(muscadet_obj$cnacalling$clusters)
# check original and remapped clusters
table(clusters, muscadet_obj$cnacalling$clusters)

muscadet_obj <- assignClusters(muscadet_obj, partition = 0.8, mapping = mapping)
table(muscadet_obj$cnacalling$clusters)
# check original and remapped clusters
table(muscadet_obj$clustering$clusters[["0.8"]],
      muscadet_obj$cnacalling$clusters)

# Visualize clusters on heatmap
heatmapMuscadet(
    muscadet_obj,
    partition = 0.8,
    filename = file.path("heatmap_muscadet_res0.8.png"),
    title = "Example sample | res=0.8"
)
heatmapMuscadet(
    muscadet_obj,
    clusters = muscadet_obj$cnacalling$clusters,
    filename = file.path("heatmap_muscadet_custom_res0.8.png"),
    title = "Example sample | rearranged clusters from res=0.8"
)
} # }