This functions computes cluster medoids given an SCseq object and a clustering partition. The medoids are either derived from the distance matrix or, if the slot distances is empty, from the dimensionally reduced feature matrix in slot dimRed$x using the euclidean metric.
compmedoids(object, part)
Arguments
object: SCseq class object.
part: Clustering partition. A vector of cluster numbers for (a subset of) cells (i.e. column names) of slot ndata from the SCseq object.
Returns
Returns a list of medoids (column names of slot ndata from the SCseq object) ordered by increasing cluster number.