R/pvalCorMat.R

In wolski/quantable: Streamline Descriptive Analysis of Quantitative Data Matrices

#' pvalues for correlation matrix -
#' @description
#' Benjamin Hochberg - adjusted
#' tests for significance of correlations among rows.
#' Transpose your matrix if you want to have it among columns
#' @param x data matrix
#' @param alternative (see cor.test)
#' @param method (see cor.test)
#' @export
#' @examples
#'
#' mat = matrix(rnorm(10*20),ncol=10)
#' res = pvalCorMat(mat)
#' image(res$pval)
#' image(res$cor)
pvalCorMat = function(x, alternative = "two.sided", method = "spearman"){
  resP = matrix(0,nrow=dim(x)[1],ncol=dim(x)[1])
  corr = resP
  res2 = resP
  for(i in 1:dim(x)[1])
  {
    for(j in 1:dim(x)[1])
    {
      if(i < j)
      {
        test = cor.test(x[i,],x[j,],alternative,method)
        resP[i,j] = test$p.value
        corr[i,j] = test$estimate
      }
    }
  }
  ## adjust p values
  res2[upper.tri(res2)] = (p.adjust(resP[upper.tri(resP)],method="BH"))
  res2[lower.tri(res2 , diag=TRUE)] = 1

  res2[lower.tri(res2)] = t(res2)[lower.tri(res2)]
  corr[lower.tri(corr)] = t(corr)[lower.tri(corr)]
  return(list(pval = res2, cor = corr))
}