R/fdrSig.R

In Mercedeh66/mirTarRnaSeq: mirTarRnaSeq

## Written by Mercedeh Movassagh <mercedeh@ds.dfci.harvard.edu>, Aug 2020

#' @importFrom stats p.adjust
NULL

#' fdrSig Ruturns FDR significant miRNA/mRNA predictions
#'
#' This function performs FDR correction on the p_values generated by the runModels function list.
#' @param RMObj The output of runModels
#' @param value The FDR value default is 0.1
#' @param method The p-value adjustment method default is fdr. It could be either of the following "holm", "hochberg",
#' "hommel","bonferroni", "BH", "BY", or "fdr".
#' @return A list of FDR corrected p vlaues, annova, and significance for each gene and the miRNA/s of interest
#' @export
#' @keywords p_adjust, correction
#' @examples
#' models <- runModels(Combine, geneVariant, "ebv-mir-bart9-5p")
#' x <- fdrSig(models, value = 0.1, method = "fdr")
fdrSig <- function(RMObj, value = 0.05, method = "fdr") {
    p_adjusted <- apply(RMObj$pvalues, 2, p.adjust, method = method)
    is_significant <- apply(p_adjusted, 1, function(x) any(is.finite(x) & x < value))
    sig_models <- RMObj$all_models[is_significant]
    sig_genes <- RMObj$genes[is_significant]
    sig_pvalues <- RMObj$pvalues[is_significant, , drop = FALSE]
    sig_p_adjusted <- p_adjusted[is_significant, , drop = FALSE]
    # add calculated values to original object and return
    RMObj$p_adjusted <- p_adjusted
    RMObj$is_significant <- is_significant
    RMObj$sig_models <- sig_models
    RMObj$sig_genes <- sig_genes
    RMObj$sig_pvalues <- sig_pvalues
    RMObj$sig_p_adjusted <- sig_p_adjusted
    return(RMObj)
}