This is also called the inverse normal transformation. We rank the data in x, divide by n-1, and take quantiles from the normal distribution using qnorm.
Usage
rankZ(x)
Arguments
x
Numeric vector of values to be transformed.
Value
Numeric vector with a normal distribution.
Details
We often use this when there are hundreds or thousands of phenotypes. This allows us to calculate permutation derived thresholds once and use them for all phenotypes.