Generates a cluster object for an scMethrix object

Generates a cluster object for an scMethrix object

cluster_scMethrix(
  scm = NULL,
  dist = NULL,
  assay = "score",
  type = c("hierarchical", "partition", "model"),
  colname = "Cluster",
  n_clusters = 1,
  verbose = TRUE,
  ...
)

Arguments

scm	scMethrix; Input scMethrix object. If this is specified the distance matrix will be a generic spearman.dist distance
dist	dist; Optional. A distance matrix generated for an assay. Will use default paramaters for get_distance_matrix.
assay	string; The assay to use. Default is 'score'
type	string; The type of distance metric to use. Available options are 'hierarchical', 'partition', "model". An arbitrary cluster function can be used, and must return a named vector containing integers representing the cluster membership (e.g. c(C1=1,C2=1,C3=1,C4=2)).
colname	string; the name of the colData column that contains the cluster information
n_clusters	integer; the desired number of clusters. This is ignored for model-based clustering
verbose	boolean; flag to output messages or not
...	Additional parameters for the clustering functions

Value

An scMethrix object

Details

Enables multiple methods of clustering to classify samples in an scMethrix object. Either an scMethrix object or a dist object must be provided for clustering.

See also

get_distance_matrix() for distance metrics, hclust() for heirarchical clustering, kmeans() for partition clustering, Mclust() for model clustering

Examples

data('scMethrix_data')
scMethrix_data <- impute_regions(scMethrix_data)
#> Starting imputation...
#> Imputed in 0.10s
dist <- get_distance_matrix(scMethrix_data,assay = "impute")  

# For a generic clustering function
# The function must return a named vector of integers
fun <- function (dist) {
    fit <- hclust(dist, method="ward.D")
    fit <- cutree(fit, k=2)
    return(fit)
    }

fun(dist) # Example of arbitrary function output
#> C1 C2 C3 C4 
#>  1  1  1  2 
colData(cluster_scMethrix(scMethrix_data, dist = dist, type = fun))
#> DataFrame with 4 rows and 2 columns
#>       Sample   Cluster
#>    <integer> <integer>
#> C1         1         1
#> C2         2         1
#> C3         3         1
#> C4         4         2