R/makeNet.R

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

#' @title makeNet
#' @description Creates an igraph object from filtered miR-mRNA interactions.
#' Resulting list can be used to display an internal R miR-mRNA interaction
#' network.
#' @param MAE MultiAssayExperiment to store output from makeNet. It is
#' recommended to use the same MAE which stores output from matrixFilter.
#' @param filt_df Filtered miR-mRNA interactions produced by the matrixFilter
#' function. This should be stored as an assay within the MAE used in the
#' matrixFilter function.
#' @return A list of igraph data which represent miR-mRNA interactions
#' filtered from the input data, wikipathway of choice and database mining,
#' This list is input for quickNet. Output will be stored as metadata in the
#' input MAE.
#' @export
#' @importFrom igraph graph_from_data_frame
#' @usage makeNet(MAE, filt_df)
#' @examples
#' Filt_df <- data.frame(row.names = c("mmu-miR-320-3p:Acss1",
#'                                      "mmu-miR-27a-3p:Odc1"),
#'                       corr = c(-0.9191653, 0.7826041),
#'                       miR = c("mmu-miR-320-3p", "mmu-miR-27a-3p"),
#'                       mRNA = c("Acss1", "Acss1"),
#'                       miR_Entrez = c(NA, NA),
#'                       mRNA_Entrez = c(68738, 18263),
#'                       TargetScan = c(1, 0),
#'                       miRDB = c(0, 0),
#'                       Predicted_Interactions = c(1, 0),
#'                       miRTarBase = c(0, 1),
#'                       Pred_Fun = c(1, 1))
#'
#'MAE <- MultiAssayExperiment()
#'
#'MAE <- makeNet(MAE, Filt_df)
makeNet <- function(MAE, filt_df){

    if (missing(MAE)) stop('MAE is missing. Add MAE. This will store the output of makeNet Please use matrixFilter first.')

    if (missing(filt_df)) stop('filt_df is missing. Add filtered miR-mRNA interaction dataframe. Please use the matrixFilter function first. The output of matrixFilter should be stored as an assay within the MAE used in the matrixFilter function.')

    metadata <- `metadata<-` <- NULL

    # Double amount of rows
    df <- rbind(filt_df, filt_df)

    # Add new column call id
    df$id <- "id"

    # names should be as.class
    genes <- data.frame(genes = c(as.character(filt_df$miR),
                        as.character(filt_df$mRNA)))

    genes$id<- as.integer(factor(genes[,1]))

    df$id <- paste("s", genes$id, sep = "")

    rownames(df) <- NULL

    # Identify the total number of rows (miR-mRNA interactions)
    halfway <- max(as.integer(rownames(df))/2)

    # Clarify the nodes
    nodes <- data.frame(id = df$id,
                        genes = c(as.character(df$miR[seq_len(halfway)]),
                        as.character(df$mRNA[seq_len(halfway)])))

    nodes$genetype <- "mRNA"

    number_miRs <- max(as.integer(rownames(nodes))/2)

    for (i in seq_len(number_miRs)) {

        nodes$genetype[i] <- "miR"
    }

    # Clarify the links
    links <- data.frame(from = nodes$id[seq_len(halfway)],
                        to = nodes$id[-c(seq_len(halfway))],
                        Databases = filt_df$Pred_Fun,
                        Correlation = filt_df$corr,
                        type = "hyperlink")

    nodes <- nodes[! duplicated(nodes$genes),]

    # Formalise nodes and links using igraph
    net <- igraph::graph_from_data_frame(d=links, vertices=nodes,
                                         directed=TRUE)
    # Store net in MAE object
    metadata(MAE)[["net"]] <- net

return(MAE)
}