Call differentially expressed (DE) genes by Clipper algorithm

callDE takes two vectors representing each gene's significance. They are usually p-values from the target data and synthetic null data.

Usage

callDE(
  targetScores,
  nullScores,
  nlogTrans = TRUE,
  FDR = 0.05,
  contrastScore = "diff",
  correct = FALSE,
  threshold = "DS",
  ordering = TRUE,
  nCores = 1
)

Arguments

targetScores

A named numeric vector of the DE scores from the target data, e.g., the p-values between two clusters from the real data.

nullScores

A named numeric vector of the DE scores from the synthetic null data, e.g., the p-values between two clusters from the null data.

nlogTrans

A logical value. If the input scores are p-values, take the -log10 transformation since Clipper require larger scores represent more significant DE. Default is TRUE.

FDR

A numeric value of the target False Discovery Rate (FDR). Must be 'diff' or 'max'.

contrastScore

A string value of the way to construct contrast scores. The choice can be

correct

A logical value. If TRUE, perform the correction to make the distribution of contrast scores approximately symmetric. Default is FALSE.

threshold

A string value of the threshold method. Must be "BC" or "DS". Default is "DS"

ordering

A logic value. If TRUE, order the genes in the returned table by their significance. Default is TRUE.

Value

A list of target FDR, DE genes, and the detailed summary table.

Details

This function constructs the contrast scores by taking the difference between target DE scores and null DE scores.

Examples

targetScores <- runif(10000)
nullScores <- runif(10000)
names(targetScores) <- names(nullScores) <- paste0("Gene", 1:10000)
res <- callDE(targetScores, nullScores, correct = FALSE)