R/returnCluster.R

In Krutik6/TimiRGeN: Time sensitive microRNA-mRNA integration, analysis and network generation tool

#' @title returnCluster
#' @description Retrieves information about which wikipathways fitted best
#' to a specific cluster. This function is to be used after quickFuzz.
#' @param MAE MultiAssayExperiment which will store the output from
#' returnCluster. It is recommended to use the same MAE which stores output from
#' the createClusters function.
#' @param clusterData A dataframe which contains cluster-pathway fit scores
#' and is stored as an assay within the MAE used in the createClusters function.
#' @param whichCluster Integer which should corresponds to the cluster of
#' interest.
#' @param fitCluster Integer from 0-1. How well should the pathways fit into the
#' selected cluster? Default is 0.99.
#' @return A dataframe that contains information about the pathways that
#' corresponded best with the chosen cluster. Output will be stored as an assay
#' in the input MAE.
#' @export
#' @usage returnCluster(MAE, clusterData, whichCluster, fitCluster)
#' @examples
#' MAE <- MultiAssayExperiment()
#'
#' metadata(MAE)[["e_list"]] <- e_list_mouse
#'
#' metadata(MAE)[["w_list"]] <- w_list_mouse[1:10]
#'
#' MAE <- wikiMatrix(MAE, ID_list = metadata(MAE)[[1]],
#'                   wp_list = metadata(MAE)[[2]])
#'
#' MAE <- turnPercent(MAE = MAE,
#'                    wikiMatrix = assay(MAE, 1))
#'
#' MAE <- createClusters(MAE, method = "c",
#'                     percentMatrix = assay(MAE, 2),
#'                     noClusters = 2, variance = 0.99)
#'
#' MAE <- returnCluster(MAE, clusterData = assay(MAE, 3), whichCluster = 1,
#'                      fitCluster = 0.5)
returnCluster <-function(MAE, clusterData, whichCluster, fitCluster = 0.99){

    if (missing(MAE)) stop('MAE is missing. Add MAE object. Results from returnCluster will be stored within this MAE. Please use createClusters first.')

    if (missing(clusterData)) stop('clusterData is missing. Add dataframe which has cluster-pathway fit scores. Please use the createClusters function first. clusterData should be stored as an assay within the MAE used in the createClusters function.')

    if (missing(whichCluster)) stop('whichCluster is missing. Add integer which represents the cluster that is of interest.')

    X <- as.data.frame(clusterData)

    # Remove pathways from cluster which have a fit score lower than fitCluster
    singlecluster <- X[which(X[,whichCluster] > fitCluster),]

    # Make a unique name for the files
    a <- "Cluster:"

    cl <- whichCluster

    b <- "_fit:"

    fit <- fitCluster

    MAE2 <- suppressWarnings(suppressMessages(MultiAssayExperiment(list(x = singlecluster))))

    # Paste unique name
    names(MAE2) <- paste0(a, cl, b, fit)

    MAE <- suppressWarnings(c(MAE, MAE2))
return(MAE)
}