Jul 7, 2008

Observed and Expected p-values

You have tested hundreds or thousands of hypothesis in a microarray experiment or a large genotyping association study and need to know if you have more significant results than expected by chance alone.

Under the null, the distribution of p-values is uniform. You can plot the observed cumulative distribution of p-values against the expected under the null. The function qqpval will help you. I hope you get something better than that:

No comments: