R/runDistMat.R

In navinlabcode/copykit: CopyKit

#' Run distance matrix calculations
#'
#' Performs distance matrix calculations that can be downstream used for
#' hierarchical clustering or phylogenetic analysis. Uses \code{amap::Dist()}
#' in order to parallelize distance calculations.
#'
#' @author Darlan Conterno Minussi
#'
#' @param scCNA scCNA object.
#' @param metric distance metric passed to calculate the distance matrix.
#' @param n_threads Number of threads used to calculate the distance matrix.
#' Passed to `amap::Dist`.
#'
#' @return A distance matrix in the slot \code{distMat} from scCNA object.
#' Access the distance matrix with: \code{distMat(scCNA, withDimnames = TRUE)}
#' @export
#'
#' @examples
#' copykit_obj <- copykit_example_filtered()[,1:10]
#' copykit_obj <- runDistMat(copykit_obj)
runDistMat <- function(scCNA,
                       metric = "euclidean",
                       n_threads = 1) {
    # cores check
    if (n_threads < 1) {
        n_threads <- 1
    }

    message("Calculating distance matrix with metric: ", metric)
    message("Using ", n_threads, " cores.")

    seg_data <- t(segment_ratios(scCNA)) %>%
        as.data.frame()

    dist_mat <-
        amap::Dist(seg_data,
            method = metric,
            nbproc = n_threads
        )

    distMat(scCNA) <- dist_mat

    message("Access distance matrix with copykit::distMat()")
    message("Done.")

    return(scCNA)
}