Interestingly, the two data sets share 8 significant pathways; we used Monte Carlo simulation to assess the significance of sharing and found the P-value is less than 1.0E-6. To examine whether our method detects biologically relevant pathways or random combinations of genes, we permutated genes and generated 215 random pathways for both EA and AA data sets; Our method only detected 6 significant pathways (P <= 0.001) in EA and 3 in AA, and none of them is shared, indicating that the number of significant pathways detected by our methods is more than what expected by chance, and those significant pathways are likely to be biologically relevant. The list of the 8 replicated pathways are shown in Table 1, together with their nominal P-values, the gene set sizes and the SNP set sizes associated with each pathway, and the full list of 22 significant pathways in EA and 11 significant pathways in AA are shown in Additional file 2.
