R/main.R

In pouryany/pathwayDisturbance: Pathway enrichment analysis of genomic using Causal Disturbance Analysis

#' Calculates enrichment analysis of KEGG pathways based on causal disturbance
#' model. For more information, see publication doi>10.1145/3107411.3107488
#'
#' @param deIDs  a list of differentially expressed genes
#' @param allIDs list of all genes in the experiment
#' @param iter   Iteration round for sampling causal disturbance
#' @param alpha  Decay factor for Source/Sink centrality, default 0.1.
#' @param beta   Relative importance of sink versus source
#' @param statEval 0 for summation of source/sink centrality 1 for product
#' @param fdrMethod choice of false disovery rate method. inherited from p.adjust
#' @param verbose printing pathway processing report
#' @return list containing pathway names, number of pathway nodes and edges,
#'         Over-representation P-values, size of differentially expressed genes
#'         in the pathway, and causal disturbance
#'
#'
#'
#' @importFrom magrittr %>%
#'
#' @export
#'
#'
causalDisturbance <- function(deIDs, allIDs, iter = 2000, alpha = 0.1,
                              beta =1,statEval = 1 , fdrMethod = "BH",
                              verbose =F){


    len     <- length(pathways.collection.names)
    res     <- vector ("list", length = len)

    for ( i in 1:len ) {
        res[i] <- list(unlist(processPathway(pathways.collection[[i]],
                                             pathways.collection.names[[i]],
                                             deIDs, allIDs,iter,
                                             alpha,beta,statEval )))

       if(verbose) { print(cat("Pathway done \n pathway name:",
                     pathways.collection.names[[i]]))
                   }
    }

    res <- as.data.frame(res)
    colnames(res)  <- names(pathways.collection.names)
    rownames(res)  <- c("Name","nodes","edges","P_ORA",
                        "No. DE","P_SSC", "causal Disturbance")

    res <- as.data.frame(t(res))

    res.clean        <- res[!is.na(res$`P_SSC`),]
    res.clean$cadia  <- p.adjust(as.numeric(as.character(
                                            res.clean$`causal Disturbance`)),
                                 method = fdrMethod)

    res.clean$ORAFDR <- p.adjust(as.numeric(as.character
                                (res.clean$P_ORA)),method = fdrMethod)

    res.clean$KEGGID <- stringr::str_sub(rownames(res.clean), end = -5)

     rownames(res.clean) <- NULL
    res.clean <- res.clean[order(res.clean$cadia),]

    res.clean[,c(4,6,7,8,9)] <- mapply(as.character,res.clean[,c(4,6,7,8,9)])
    res.clean[,c(4,6,7,8,9)] <- mapply(as.numeric,res.clean[,c(4,6,7,8,9)])
    res.clean[,c(4,6,7,8,9)] <- mapply(formatC,res.clean[,c(4,6,7,8,9)],
                                    MoreArgs = list(format = "e", digits = 3))

    res.clean      <- dplyr::as_data_frame(res.clean) %>%
                      dplyr::mutate_at(.,.vars = c(4,6,7,8,9) , as.numeric)
    res.clean$Name <- as.character(res.clean$Name)

    return(res.clean)
}