Benchmark of false positive calling. Shown are estimates of P(P value<0.01) under the null hypothesis. The FPR is the number of P values less than 0.01 divided by the total number of tests, from randomly selected comparisons of five vs five samples from the Pickrell et al.  dataset, with no known condition dividing the samples. Type-I error control requires that the tool does not substantially exceed the nominal value of 0.01 (black line). EBSeq results were not included in this plot as it returns posterior probabilities, which unlike P values are not expected to be uniformly distributed under the null hypothesis. FPR, false positive rate.