Hi everyone,
I've calculated PBS values for my SNPs and now I would like to test their significance using the qvalue
package in R.
This package asks you for either a list of p-values or (if you don't have them) to calculate them from a null distribution of your data with the function empPvals
. My questions are the following:
1) How can I calculate the null distribution of my PBS data? In the package example, two-sample t-test labels are permuted between two different groups to generate this distribution. However, in PBS there's nothing you can "permutate", meaning, each SNP will have its own PBS value.
2) Would it be a good approach to rank my SNPs by PBS values and then divide by the total number of SNPs (to get 0 to 1 values) and consider these the empirical p-values?
Thanks a lot in advanced!